The efforts required to establish a framework for high-volume, accurate annotation are substantial. I believe that it is important that we reflect on what we have learned about the factors that determine productivity. So, what have we learned from the project?

First, subsystem-based annotation is the key to accuracy. While there are certainly numerous efforts still focusing on annotation of a single genome, the recognition that comparative analysis is the key to everything, and that focusing on the variations of a single component of cellular machinery as they are manifested over the entire collection of existing genomes is the key to accuracy, are both widely accepted principles at this stage. Manually-based subsystem creation and maintenance is the rate-limiting component of successful annotation efforts, and the factors that constrain this process are at the heart of the matter. We have understood this for some time now.

I am going to argue a new position in this short essay:

1. There are three distinct components that make up our strategy for rapid accurate annotation: subsystems-based annotation, FIGfams as a framework for propagating the subsystems annotations, and RAST as a technology for using FIGfams and subsystems to consistently propagate annotations to newly-sequenced genomes.
2. These three components form a cycle (subsystems => FIGfams => RAST technology => subsystems). This cycle creates a feedback that rapidly accelerates the productivity achievable in all three components. Further, failure in any of these components impairs productivity dramatically in the others. Understanding this cycle will be the key to supporting higher productivity in subsystem maintenance and creation.
3. To understand the dependencies, we need to consider each of the components:
   • The key to accurate FIGfam creation and maintenance is to couple it directly to subsystem maintenance. Once the initial release of the FIGfams was created, updating them occurs automatically based on changes in the subsystem collection. Thus, FIGfams are automatically split, merged and added as the subsystem collection is maintained. There remains one area of substantial cost in FIGfam development -- creation of family-dependent decision procedures that are occasionally required to achieve the required accuracy. At this point we have approximately 10,000 subsystem-based FIGfams, although the overall collection contains over 100,000 families (the majority containing only 2-3 members).
   • RAST has a central dependency on FIGfams for assertion of function to newly-recognized genes. In this sense, the main dependency of RAST is on the FIGfam collection. The more accurate the FIGfams and their associated decision procedures, the more accurate the assignments of function made to genes in genomes processed by RAST.
   • Finally, the central costs of maintenance of subsystems are cleaning up errors in existing subsystems (often indicated by multiple genes having the same function) and by adding new genomes to existing subsystems. Once a subsystem has reached an acceptable level of accuracy (and many are not there yet), the central cost is integration of new genomes after annotation by RAST. The speed with which new genomes can be added depends on how well RAST assigns gene function (and, secondarily, on how accurately these RAST-based annotations can be used to infer operational variants of subsystems).
   The main costs of increasing the speed and accuracy of annotations split into two categories: those relating to maintenance of existing subsystems, and those relating to generation of new subsystems. The maintenance costs are containable, if the cycle is established and functions smoothly. Otherwise, I suspect they inevitably grow rapidly.

I have argued that the costs in achieving rapid, accurate annotations is limited by the rate at which subsystems can be maintained and created. I place the maintenance ahead of creation at this stage. As the collection grows (it now contains over 600 subsystems with over 6800 distinct functional roles), costs of maintenance will tend to dominate. The creation of new subsystems will always be a critical activity, but each new subsystem will impact smaller sets of genomes as we "move into the tail of the distribution".

The costs relating to subsystem maintenance, which will quickly dominate, depend critically on how smoothly the cycle I described functions. We have just established the complete cycle, which is arguably the major achievement of 2007.

The two central costs that cannot be avoided will be creation of FIGfam-dependent decision procedures and the creation of new subsystems. We are currently beginning focused efforts to increase annotator productivity relating to both activities. The required technology in these cases is less dramatic and relates to development of better tools (to support a set of possible decision procedures in the case of FIGfams, and to resolve inconsistencies in subsystem initiation/expansion).

More Effective Integration of Existing Annotation Efforts

In the section above, I reflected on the cycle that we shall depend upon for supporting increased volume and accuracy of our own efforts. Other groups are certainly experimenting with their own solutions, and in some cases with clear successes. I have no desire to rate these competing efforts. We have a new year coming, and I sincerely believe that cooperative activity is the key to enhanced achievements by everyone. However, effective cooperation is often elusive. I think that we have put in place an extremely important mechanism for making cooperation much easier, and the benefits more compelling.

Anyone working for one of the main annotation efforts realizes that it is not easy to really benefit from access to the annotation efforts of other groups. The efforts required to characterize discrepancies between local annotations and those produced externally often outweigh any benefits that result.

Two events of major importance have occurred:

1. Both PIR and the SEED Project decided to build correspondences between IDs used by different annotation projects. The PIR effort produced BioThesaurus and the SEED effort produced the Annotation Clearing House. The fact that it will become trivial to reconcile IDs between the different annotation efforts will undoubtedly support rapid increases in cross-linking entries. The SEED is working with UniProt to cross-link proteins from all of our complete genomes, and I am sure similar efforts are happening between the other major annotation efforts.
2. Within the Annotation Clearing House, a project to allow experts to assert that specific annotations are reliable (using whatever IDs they wish) has been initiated. This has led to many tens of thousands of assertions that specific annotations are highly reliable. PIR is preparing a list of assertions that they consider highly reliable, and both institutions are making these lists openly available.

To see the utility of exchanging expert assertions in a framework in which it is easy to compare the results, let me describe how we intend to use these assertions:

1. We begin with a 3-column table of reliable annotations containing [ProteinID,AssertedFunction,IDofExpert]
2. We then take our IDs and construct a 2-column table [FIG-function,AssertedFunction]. This table gives a correspondence between each of our functional roles and the functional roles used by the expert making the assertion of reliability.
3. Then, we go through this correspondence table (using both tools and manual inspection) and split it into one set in which we believe both columns are essentially identical and a second set that we believe represent errors (either our own or those of the expert asserting reliability). We anticipate that in most cases the expert assertion will be accurate, which is what makes this exercise so beneficial to ourselves.
4. We take the table of "essentially the same" assertions and distribute it as a table of synonyms (which we consider to be a very useful resource).

We are strongly motivated to resolve differences between our annotations and high-reliability assertions made by experts. The production of the table of synonyms both reduces the effort to redo such a comparison in the future, but is also a major asset by itself. I am confident that any serious annotation group that participates will benefit, and I believe that these exchanges will accelerate in 2008 and 2009.

Summary

I have tried to explain why I think that 2007 was a watershed year. The creation of the subsystems => FIGfams => RAST => subsystems cycle was not precisely planned, but its achievement has made its value obvious. That, coupled with the growing tendency to cross-link and exchange reliable assertions will lead to rapidly improving annotations over the next 2-3 years.

Although we now probably have over a thousand sequenced genomes, we have not yet integrated that many into the SEED (and I doubt that we would have access to that many right now). However, it seems very, very likely that we will have that many by sometime in 2008, and it also seems likely that we will be in a position to provide pretty decent initial annotations. I would anticipate completing the original project next year, and it is now time to plan the next stage.

Development of the NMPDR

As we mentioned in our last newsletter, FIG received a sizable portion of a 5-year grant to build a National Microbial Pathogen Database Resource. This grant is providing a welcome stability. In addition, the NMPDR is a completely open source project, and the extensions to the existing SEED technology that are under development will benefit everyone. Many aspects of the basic design are being refined and cleaned up. The first release is only a few weeks away. While based on SEED technology, the team doing the development has introduced a number of key innovations that we will discuss in the next newsletter.

The Manifesto

A discussion has been taking place relating to the appropriate objectives of FIG and the SEED project. Now that FIG has solid funding due to its participation in a the development of the NMPDR, there is a welcome chance to re-examine goals. There are a number of ways things might develop, and as one might expect, different views are constantly being expressed. Ross attempted to formulate his view, which was posted at http://TheSEED.uchicago.edu/FIG/Html/1KG.html. The focus he advocates is largely based on annotations: success should be measured by FIG's ability to advance the development of subsystems and detailed encodings of metabolism (in the form of stoichiometric matrixes).

The SEED Developers Meeting in Chicago in October

The SEED Developers meeting was held on Oct 24-25. It focused on preparing a distribution version of a merged environment that would support both the SEED and GenDB. Teams at both Bielefeld and Argonne National Lab put substantial effort into merginjg the systems. A prototype was produced, DVDs were created (shortly after the meeting), and the result was shown at Supercomputing 2004, which occurred in early November. The merged system is not completely distributable yet for two reasons:

1. The installation scripts are not solid. We can manually do an install by working through the steps, but the details for making things work properly in arbitrary environments is taking time.
2. It became clear that we wanted to include a number of genomes in the GenDB side that had all of the precomputed data needed to support analysis (rather than making users begin by doing a fairly substantial amount of computation to prepare their genomes of interest).

The effort to make both systems available in an integrated environment will continue during the first quarter of 2005. If anyone really wants a copy right away for production use (as opposed to just evaluation), we will make an effort to help you install it. However, the major release scheduled for the end of the first quarter will include hundreds of genomes already preloaded into the GenDB component, tens of thousands of newly-called genes, and on the order of a hundred additonal genomes. We sincerely believe that it will be worth the wait.

Workshops, Tutorials, etc.

We held tutorial/workshops at MIT and in Mexico, and things went extremely well. We have another scheduled in early March at the University of Florida, and Andrei will be giving one in the Netherlands in May. We have been reconsidering the most effective way to spread the technology, given that everyone is already hugely over commited.

One observation has been that, in the few cases in which the SEED technology for annotation of genes and subsystems has actually been used in classrooms, we got more benefit for our efforts. Most notably, we had a really productive experience helping introduce the technology in Bernhard Palsson's graduate class at UCSD. The students took it very seriously, they built a number of well-done subsystems, and it established the foundation for ongoing use and collaboration.

The position that is emerging is that advancing the effort to produce accurate, well supported subsystems should be the main goal; all tutorial efforts would be judged on the odds that the expended effort would advance that goal. Hence, tutorials given to experts planning on using the technology to support development of reveiws or in their own research would get highest priority, graduate classes would get next highest priority, etc.

New Releases, etc.

The effort required to add hundreds of genomes at an increasing pace is draining. Just trying to help people make new installations and update their old ones is time consuming. It is becoming very clear that we need to establish a strategy that scales, and we need to implement it very quickly. The current plan is really pretty exciting. The key points are as follows:

1. The major release that we will make about the end of the first quarter will be the "first and only more-or-less official release of the SEED/GenDB data". Code will be constantly updated and installed over the network as it is done now. However, we will stop constructing "current releases".
2. Rather than new releases of an ever-increasing collection of genomes, we will support incremental addition of single genomes. Users can download genomes, similarities, improved gene calls, etc. from the clearinghouse. Users will download and install genomes the way they now download and install subsystems.
3. A number of participants will be helping to prepare genomes for addition, and they will all be depositing them into the clearinghouse.
4. For new participants, we will provide the single release, and we will implement the ability to "clone" and existing SEED/GenDB system. This allows a new user to acquire a "clone" from someone who has made the investment to download whatever genomes are desired, install whatever subsystems they trust, etc. It also frees us from having to worry about how to help people get started.

There are numerous aspects of this strategy that have not been properly implemented yet. We will do our best to have everything fully functional by the big release.

The More Important Release: the Subsystems

The big release will include not just the SEED/GenDB integration, but an initial set of subsystems, as well. These represent the output of the "prototyping and evaluation" stage. There will be on the order of 100-150 subsystems in differing stages of development. Some were done by experts, have lovely diagrams, and reflect major amounts of effort. Many were done by enthusiastic participants with less knowledge and they reflect it. There are many ways that one might measure the utility of this initial batch. We believe that it represents a major milestone. Much of the next two months will be spent organizing and doing quality control for this initial release.

Well, that is about it for now. There are probably many things that we forgot to cover, but we did hit the major topics. We will try to get out another newletter next month, but it is not clear that we will take the time to do so before "the big release", which should be about the start of March. In any event, we wish you well and hope you prosper,

the team at FIG

The reason we had not originally planned on doing the whole thing at ANL was based on worries about being able to get people into the lab (we must get people cleared by ANL security). We have checked, and it appears that everyone (that we know about) coming to the second half can easily be cleared. So, it is critical that, if you have not discussed your attendance with us, but you would like to come, please contact us immediately. If it emerges that there are people who would like to come, but cannot get clearance, we will run a small parallel version at the University of Chicago (or, we will help them get set up for a second tutorial that we will be holding at MIT in Boston on Nov 1-2).

So, that is the important message: PLEASE MAKE SURE THAT WE KNOW IF YOU ARE COMING AND MAKE SURE THAT WE HAVE GOTTEN CLEARANCE FOR YOU.

Agenda for the Developers Meeting (Oct 25-26)

The SEED Developers Meeting will be held at Argonne National Lab on Oct 25-26. There are good cheap hotels available (on site for about $65/night; contact Veronika or Cheryl Zidel).

These meetings are intended to be largely free format with intense exchanges between individuals extending the SEED. Inevitably small subgroups will form and work on specific technical issues. However, we will consider the following to be a draft agenda:

October 25

Building 221, room A216, Mathematics and Computer Science Division, Argonne National Lab (wireless connections are available, so bring your laptops)

We must stress that not everyone will be attending these discussions. One group will organize itself to focus on the development of the distribution DVDs. Members will come and go. It will (intensionally) be a somewhat unstructured, hopefully intensly productive, exchange.

October 26

The three days that follow (Oct 27-29) are intended to be less technical, more tutorial, but also somewhat unstructured. We anticipate that many users will come only for these three days.

October 27: What Can One Do With the SEED?

October 28

October 29

That is approximately what we plan to do.

Hope you can make it.

The next SEED Developers Meeting will be in Chicago during the week of October 24-28. The meeting will be split between Argonne National Lab (Oct 24-25) and the University of Chicago (October 26-28).

The initial period at Argonne will focus on preparing a distribution version of the GenDB/SEED merged environment. This should produce a set of DVDs with both GenDB and the SEED in the distribution. The merged environment is built on technology developed at the University of Bielefeld in Germany and the combination should offer capabilities that are attractive to many groups. In particular, it should support straightforward editing of genes (adding CDSs, deleteing them, and changing start locations), as well as comparative capabilties. This part of the meeting will be quite technical. Anyone planning on attending should contact Ross Overbeek (Ross@TheFIG.info) or Mike Kubal (mkubal@mcs.anl.gov) till September 15th. Entrance to Argonne is restricted, and we need to apply for guest passes for visitors well in advance (especially foreign nationals).

The second part of the meeting will be held at the University of Chicago and will focus on the Project to Annotate 1000 Genomes (i.e., subsystems annotation). This will include both a tutorial for people that wish to understand how to annotate using the SEED, as well as an exchange of views and status by people already quite active in the effort. For those wishing to attend this part, please contact either Mike or Ross, and feel free to do so right up to the meeting.

The Subsystems Effort: How it is Shaping Up

The subsystems effort is beginning to take shape. It now involves a number of people, some real experts and some relatively inexperienced. Some people work on the SEED server at the University of Chicago (http:TheSEED.uchicago.edu/FIG), but more work on their own personal systems. Periodically, they deposit versions of their analysis on a server from which others can download as desired (this is called "publishing" the subsystem). Any user of the SEED can download any or all of these deposited subsystems.

A Wiki is used to coordinate who is working on what (http://www-unix.mcs.anl.gov/SEEDWiki/moin.cgi/SubsystemBulletinBoard). A Forum (http:subsys.info) has been set up as a framework for posting notes of general interest, predictions, and general comments. The forum is intended to become a vehicle for opening up the effort to an unlimited set of potential users.

We have found that there has been a bit of effort required to get people acquainted with use of the Wiki and the Forum, and use of the software to support development of a subsystem does require a tutorial. There are basic inadequacies in the tools at all levels, and this can prove to be quite frustrating.

However, these last three weeks (vacation time seems to come to an end) have seen an amazing amount of progress, and things are now beginning to take off. A number of subsystems that cover thousands of assignments have been published in these last few weeks, and we know of many more in progress. We are very enthusiastic about this point.

The iBook: a Minor Technical Note

Sveta Gerdes recently joined FIG and has begun working on encoding subsystems. Due to some lovely technical work by Bob Olson and Ed Frank, the entire SEED can now be run on a laptop without an external drive. It does consume 20-30 gigabytes of the internal disk. Sveta now uses an iBook (basic cost of about $1499 + $129 for and additional 512 MB of internal memory) that supports a full version of the SEED on the internal disk.

Leading to a Proposed Annotation Environment

This leads us to briefly discuss the basic working environment that we are proposing for annotation teams:

1. A central SEED server is optional, but is often useful for two reasons: it can be used to support people that have only really inexpensive laptops or desktop computers, and it can be used as a framework for offering access to the world at large. Most annotation efforts will include one, and it can be run on anything from a $1500 Linux box to a $4000 Mac G5.
2. Some members of the annotation team will have moderate laptops ($1600-3000 machines) running local copies of the SEED. Others will use personal computers to work over the network on the central SEED server.
3. Function assignments and annotations are exchanged between individual SEED versions of members of the team (and the central server, if it exists) each evening.

The currently available peer-to-peer tools in the SEED are not adequate to effectively support this style of synchronization. We consider it a high priority to get them to this point in the next month or so.

This leads us to a central tenant of the proposed environment: assignments and annotations are the output of an annotations effort, and there must be an external representation that defines exactly what is meant by their content. In the presence of such a definition, individual users can use whatever annotation software they wish, as long as the results can be integrated and synchronized with those of the entire team.

Once environments are established that contain multiple annotation systems, all synchronizing their results on a daily basis, systems will evolve to take over niches (sorry, perhaps the level of enthusiasm is clouding our judgement here...).

Tutorials

Ross is planning on teaching a 3-day course in annotations and subsystems technology at UCSD on Oct 4-6.

We plan on having a subsystems tutorial and general discussion of SEED technology at MIT in Oct (probably Oct 18-19).

The cancelled tutorial at Los Alamos will be rescheduled for Sante Fe in November.

If you wish to know about any of the details, please check the FIG Forum or contact Veronika (Veronika@TheFIG.info).

The Funding Situation

During FIG's first year, the funding situation might reasonably be characterized as somewhat precarious. We survived due to a few relatively small contracts, and some donations from people and institutions. We are deeply grateful to those who helped support our goals during that period. The plan was always to establish a basic open source platform and then to use it to secure portions of grants that would build upon those capabilities. Numerous grants were submitted in which FIG was included.

Finally, in August FIG signed a subcontract finalizing its participartion in a 5 year project led by PI Rick Stevens and Co-directed by Ross to build a National Microbial Pathogen Database (http://www.niaid.nih.gov/dmid/genomes/brc/awards.htm). This is an $18M grant made to the University of Chicago. We intend to participate in a number of grants; in some we play a very minor role helping to develop independent teams based on SEED technology, and in others we will play a more active role. The SEED technology will prove useful in a number of the emerging application areas ranging from pathogen databases to analysis of environmental samples, and we encourage groups to consider basing long-range developments on it.

We have just returned from the SEED Developes meeting held in Bielefeld, Germany on July 5-9. It was a remarkably good experience. There were three distinct events going on:

1. On July 5-6 a tutorial was held on how to annotate subsystems using the SEED. Participants from Germany, France, Denmark, the Netherlands, Poland, and Russia were there. It was a productive exchange, and it seems very likely that a number of expert biologists will complete their versions of subsystems and publish them to the clearinghouse.
2. On July 5-9, Ross taught a course in use of the SEED to a number of students and post-docs from Bielefeld. It has become apparent that the right way to teach a course about use of the SEED is to focus on development of subsystems. Before we had the subsystems-encoding component of the SEED operational, we taught several courses in use of the SEED for comparative analysis, and we sincerely believe they were quite successful. However, it is now clear that developing a 4-lecture course with assignments focusing on the topic of encoding subsystems is needed badly; this is clearly the emphasis that should exist in a "short module" that could be included in standard courses in molecular biology, biochemistry, and microbiology.
3. Finally, on July 7-9 a major effort was initiated to link the SEED with GenDB. GenDB is a serious annotation system developed by the team at Bielefeld. A substantial amount of effort and resources have gone into a system that offers many of the features needed to carefully annotate prokaryotic genomes. Many of these features are missing in the SEED. On the other hand, the SEED offers many services in comparative analysis which are missing in GenDB. The systems complement one another in several ways. Hence, we decided to consider the question "Can they easily be coupled?". The team at Bielefeld had already spent time developing a detailed protocol for linking major components of their system (allowing almost independent development, with each component having access to the data maintained by other components). In short, they had thought about the general issues that inevitably come up in such exercises. Within the 3-day session,
1. GenDB was installed at Argonne (the SEED was already running on the Bielefeld server under Solaris).
2. The access protocols were rapidly installed by the Bielefeld team, which allowed GenDB to access data stored in the SEED. This allowed a user running GenDB to "import" one of the many genomes stored in the SEED into GenDB. It is hard to convey the level of enthusiasm that led to this rapid integration. Among other things, this will allow users to perform detailed editing of gene locations (adding missed genes, adjusting starts, deleting genes that are apparently not real, etc.) This type of detailed editing will be necessary for work with many types of features (e.g., transposons and regulatory sites), and the SEED simply lacks the capabilities to do it properly at this point.
3. Detailed plans for modifying the SEED to accept updates relating to features, to smooth out the details of gene calling, and to get computational servers up and running (at both Bielefeld and Argonne) were all discussed. It seems very likely that we will be able to distribute DVDs containing both systems, which we believe will be a major step forward for support of annotation projects, within just a few months.
4. Meetings over the internet using the Access Grid were used on a daily basis to support coordinating the Argonne, FIG and Bielefeld teams. These will continue on a weekly basis.

A great deal was accomplished in a very short time. The only known failure related to Ross' promising to locate a "missing gene" in Corynebacterium glutamicum. He concedes that the group could not find one in time, but still insists that one will result from the exercises that people began. His optimism may exceed his judgement.

It was overall a truly magnificent meeting.

Subsystem Tutorials

The 2-day tutorial held in Bielefeld was completely different than the 1-day one held in Chicago. Although both events were fun and productive, the addition of the extra day allowed people to get much deeper into the topic. The plan called for people starting to work on their subsystems of interest by noon of the first day, which allowed people to confront central issues more quickly. By the end of the second day, everyone was fairly far along, and a few had relatively large spreadsheets. While it is certainly true that it takes months to accurately analyze most of these systems (if not years), it is also true that the existing functional assignments can be dramatically improved with just a day or two of effort by someone interested in participating in this project.

We now believe that annotation of subsystems needs to be the major focus of our efforts over the coming 4-6 months. We must extend the software, offer as many tutorials as possible, and actively seek experts who are willing to participate. The next subsystems tutorial will be held at Los Alamos on August 26-27. Two more are in the very loose planning stage, but we expect to be able to clarify those plans by the time we send the next newsletter.

The Subsystems Forum

Over beers, a number of us discussed the issue of recording the events that will be taking place over the next 2-3 years. If the genes implementing the central functional roles of cellular subsystems are actually worked out, it would be desirable to keep an electronic record of the key events. In addition, it would be nice to have a means of rapidly exchanging information on the status of conjectures, wet lab confirmations, and so forth. Hence, we funded the development of a "Subsystems Forum", which was started just before the Bielefeld meeting. It will be hosted at the University of Chicago. We are urging those people developing annotated subsystems to post both questions and "open problems". The URL for the forum is

or http://subsys.info

Gene Calling

The subject of gene calling was discussed at the developers meeting. The Bielefeld team has put up a server which already calls protein encoding genes and tRNAs. It can be accessed via

The general plan is as follows:

1. We will develop the tools to construct a SEED-formatted organism from the output of the Bielefeld gene calling server, including the rRNA caller developed by Niels Larsen.
2. GenDB offers a framework for examining and evaluating the relative merits of different gene calling algorithms. We will work on making it possible to easily compare the output from different systems.
3. Ralph Butler and Ross will be working on a system to recall starts for genes from a single subsystem.
4. It is important that, if we succeed in improving calls substantially that we make the results available to NCBI so that the improvements are not lost.

While the existing gene calls are excellent for some organisms, they are truly awful for others. Hopefully we can make rapid progress in improving the situation over the coming months.

Computational Servers

ANL wrote a grant in which it was proposed to (among other things) supply computational servers to support adding new organisms to the SEED. It is not clear exactly what servers are needed to support things like gene calling, adding organisms to GenDB and adding organisms to the SEED. We will work the details out over the coming few months, coordinate with efforts that already exist to provide such services, and hopefully have things running somoothly by the fall.

The major SEED development effort for the last two months has centered on annotation of subsystems. The basic goal is to develop and support the tools needed by an expert to annotate/analyze a specific metabolic or nonmetabolic subsystem over a large collection of genomes. We believe that development of these tools and the framework needed to effectively support comparative analysis is essential to straightening out the existing annotations.

The initial set of tools was released in March, and a small group of people began using them. The main objective was simply to make them reasonably reliable, to identify exactly what functionality was most needed, and to establish a framework for saving and exchanging results. By May, we had demonstrated the basic utility of the tools. In at least two cases, the subsystem annotations that were developed during the shakedown phase resulted in serious conjectures relating to "missing genes". One of these conjectures has evidently been proven wrong, but it did result in a rapid identification of the correct gene (this is a tail best told over beers).

On June 18, we had a 1-day tutorial on use of the SEED to annotate subsystems, and it was a remarkably pleasant event. We had close to 30 people attend. The main complaint was "so little time, so much to do... we need two days". Hence, in the future subsystem/SEED tutorials will be 2-day affairs. The next one will occur on July 5-6 at the University of Bielefeld in Germany, and it will be held as part of the SEED Developers Meeting (which will be July 5-9). We are planning on having one at Los Alamos in August, if schedules permit, and one in Boston in the fall. Planning on these last two needs to progress, and we will try to keep everyone posted.

There is a great deal that could be said about the significance of the subsystem annotation effort. While many SEED users and developers are more interested in annotating single genomes or using the SEED for other purposes, we at FIG view the annotation of subsystems to be the essential core of The Project to Annotate 1000 Genomes, and as such it is one of the most critical and exciting components of the SEED collabortion. We believe that this will become obvious and accepted during the next 4-6 months. Time will tell.

SEED Developers Meeting

The next major SEED event will be the meeting at Bielefeld on July 5-9. There are a number of goals for that gathering:

1. We will hold the subsystems tutorial on the first 2 days, and it appears that a number of people will attand just that component of the gathering.
2. Ross is supposed to teach a class for graduate students during the week. There will clearly be a major overlap in content between the subsystems tutorial and the class (which will, hopefully, include sections on searching for missing genes and where the technology is going).
3. On Wed-Friday, we plan on focusing on issues relating to supporting the development and maintenance of a set of communicating systems. Initially, we will focus on GenDB (the system developed at supported by the team at Bielefeld) and the SEED (the system FIG/ANL/Burnham have been developing). Both systems are open source, both groups are deeply interested in creating a joint framework that supports effective use of a loose integration of the systems, and both groups are forging ahead as quickly as possible.

The Bielefeld team has developed and is now shaking down a web service that will call genes in prokaryotic genomes. We have found this extremely useful in our efforts to produce a new release of the SEED. Ralph Butler is working on a tool to help predict more accurate start locations for CDSs, and we are slowly improving our installed genomes. Of course, we are taking RefSeq from NCBI, and they are making substantial and constant progress in cleaning up their collection. The issue of how everfything fits together, how users can install their own data on local copies of the SEED, and how exchange of data occurs will all be major topics at Bielefeld.

Funding of SEED Development

Until very recently support for development of the SEED came from a few consulting contracts obtained by FIG and by a few subcontracts that were (and are) deeply appreciated. This got FIG through last year, and it got the SEED to the point where it has demonstrable utility.

The basic model for funding SEED development goes as follows:

1. We hope to have many institutions participating. Each is responsible for getting its own funding.
2. Any institution that wishes can base projects (and proposals) on the SEED. They may, or may not, include FIG as a participant or subcontractor. No one should feel any pressure to include FIG, since the SEED technology is completely available free of charge to anyone.
3. FIG has played a leadership role until now due to the fact that Ross Overbeek did the majority of design and implementation during the first year. Argonne National Lab has begun to play a major role, the University of Chicago is building a major project upon the SEED, and it seemd likely that actual leadership of the effort will occur on a largely informal basis and be shared between a growing number of senior participants (or junior, for that matter, if they demonstrate they can do it).

FIG and the Computational Institute at University of Chicago have been awarded a sizable five-year contract to construct a National Pathogen Database. This immediately solidifies the future of the SEED effort, since that project will certainly be based upon SEED technology. It will also ensure that rapid development of new capabilities (most notably inclusion of expression, structural, and variation data) will occur. We encourage other organizations to build upon the SEED (thus, amortizing development costs over more projects).

That is all for now. We plan to write a more extensive newsletter after the Bielefeld meeting. There are a number of exciting things that are planned for that gathering, and we anticipate that there will be much to report.

]]> 2006-12-10T13:36:36Z 2004-04-29T15:38:17Z tag:www.figresearch.com,2004:/news//3.105 2004-04-29T15:38:17Z
The next meeting of SEED Developers Subsystem Annotations...
Veronika Vonstein veronika@thefig.info Newsletters
4. The next meeting of SEED Developers
5. Subsystem Annotations
]]> The Next Meeting of SEED Developers

The next meeting of the SEED Developers will be in Bielefeld, Germany on July 5-10. The meeting will be split into two parts:

July 5-6 will be devoted to Annotation of Subsystems

The main activity during these two days will be an intensive tutorial directed towards researchers that wish to annotate specific subsystems using the SEED. We will help people begin to analyze their subsystem of interest as it manifests itself in the 250+ genomes in the current SEED. This tutorial, as well as the whole topic of subsystem annotation, will be discussed in below.

A second component of these two days will involve an open discussion between researchers currently using the SEED to annotate subsystems. The topics will include

1. Which subsystems are now being analyzed?
2. How should we prioritize the work?
3. How should we be presenting the research conjectures that are already apparent from the existing efforts?

July 7-10 will be largely devoted towards linking systems

There are at least three systems that we will be working with. GenDB (the annotation system being developed at Bielefeld), the SEED, and Niels Larsen's "tree viewer". An over-simplified analysis would be that

GenDB is a "DNA-oriented" system, allowing one to identify and manipulate features on chromosomes (or contigs), the SEED is largely "protein-oriented", and Niels has focused on tools for presenting data.

The merge of the capabilities of GenDB and the SEED will be achieved (we hope) by architecting a fairly general interface for both systems (allowing GenDB to interface with other "protein-oriented" systems and the SEED to interface with other "DNA-oriented" systems). Hopefully, a user will be able to flip back and forth between environments, and work done in each environment will be communicated back to the other.

Niels' interface tools will be needed to display and browse large trees. This is particularly useful when overlaying gene sets onto functional overviews (e.g., during interpretation of microarray data, examination of complementary metabolisms between host and symbiont, and so forth). Niels is also building open source tools, so we believe that this represents a good overlap in interests. He will be visiting from Denmark.

The success of the efforts to merge the systems, get everyone familiar with the new Bielefeld server for calling genes in prokaryotes, and discuss servers for computing similarities will depend to some extent on how much time we devote in June to get ready. We are planning on having weekly discussions via conference calls (access grid communication is definitely needed and hopefully will become the basis for these meetings before to long).

It is likely that in parallel to these activities Ross, with help from his friends, will offer a tutorial in use of the SEED for students at Bielefeld. We previously had a very good experience with a 2-day intensive class at Franklin and Marshall College. This time, we plan on making it less intensive (1 hour per day) over a somewhat longer period. The basic idea would be to introduce the students to the SEED, suggest ways to find research topics, and so forth.

We really do not have an accurate idea yet who will be coming. It seems likely that there will be a number of people coming for only the first few days. In any event, please contact Alice McHardy, Folker Meyer, or Ross Overbeek if you plan on attending. Everyone will be welcome, but we do have to get a sense of who will be there in order to handle local arrangements. Alice can help you by suggesting a hotel.

Subsystem Annotations

Annotation of subsystems across many genomes is rapidly becoming a central component of the FIG collaboration. Researchers from at least six distinct institutions are now actively working on the project. This is pretty exciting, since the initial tools were released only last month.

We believe that the current tools are working well in the sense that

1. backups now occur automatically,
2. the peer-to-peer exchange of subsystems between completely different versions of the SEED (e.g., with different versions of RefSeq) seems to work well, and
3. weekly updates introduce features that are rapidly reducing the effort required to construct a reasonable spreadsheet for a subsystem.

We believe that major new tools that will significantly increase the productivity of participants should be available within two weeks. After those become available, tested, and incrementally improved, we will make the entire set of tools available to researchers on a public server (probably at the Uinversity of Chicago, initially).

The key idea of subsystem annotation is simple: a person wishing to study a subsystem using comparative analysis analyzes exactly which organisms it occurs in, which alternative variants can be distinguished, which functional roles are present in each of the organisms, and which genes implement those roles. This is roughly the raw data behind what you see in many biological review articles. By organizing the data in a form that can be supported within the SEED, it becomes much easier to analyze new genomes as they become available, to clarify outstanding uncertainties, and to curate the system over time. We believe that something like what we are now implementing will be the foundation for many, many researchers analysis over the coming years. As thousands of diverse genomes become available, this becomes the framework for a person to maintain an accurate portrait of the system or systems with which he works.

We will hold a tutorial on how to use the system in early July in Bielefeld, as we mentioned above. We will probably hold one at Argonne National Lab in Chicago before then. The tutorial will include a general overview of how to use the SEED, so we suspect that it will be of general interest. If you would be interested in attending a 2-3 day tutorial in use of the SEED and annotation of subsystems, please contact Veronika or Ross .

]]> 2006-12-10T13:36:36Z 2004-03-03T17:46:31Z tag:www.figresearch.com,2004:/news//3.108 2004-03-03T17:46:31Z
The End of the Development Stage The Initial Deployment Stage The Project to Annotate 1000 Genomes...
Veronika Vonstein veronika@thefig.info Newsletters
6. The End of the Development Stage
7. The Initial Deployment Stage
8. The Project to Annotate 1000 Genomes
]]> Well, the SEED Developers' meeting in San Diego has completed. With the start of these meetings, the SEED Project is now moving from an development stage into what might be called an "initial deployment stage". The development stage succeeded in producing a system that has some remarkable properties:
1. It does support comparartive analysis of a rich set of genomes. The nonredundant protein database used to develop similarities now has 1.92 million entries, and the entire database contains over 270 more or less complete genomes.
2. It is being used in at least two sequencing/annotation efforts as a framework to analyze the data.
3. It has been used successfully in several short courses as a framework for students to explore genomic data. We could that within two days, the central functionality could be conveyed, and students were able to focus on meaningful analysis issues. In one case, several potential research topics were exposed.
4. It is good enough to justify initiating what we consider a key FIG effort: the Project to Annotate 1000 Genomes.
5. The system can be easily installed. The process is not yet perfect, problems arise in almost every case, but we can usually have a system up with only a few minutes of human effort (the time to transmit copies over the network and to build the database is substantial, but normally requires no intervention). It runs on both Macs and Linux systems using either Postgres of MySQL.
6. It supports a rudimentary p2p update/data exchange capability. This part still needs work, but not too much. It will, like the configuration/installation process, become solid fairly rapidly as deployment begins.

We view the next stage as potentially quite exciting, and certainly pretty demanding. The most important development has been our increase in manpower: the SEED effort now includes four part-time researchers and we have increasing help from collaborators (perhaps, "participants in the project" is a better term). We are receiving help in enhancing the system from a number of individuals, and in some cases efforts have already begun to make it possible for independently developed tools to be easily integrated within the SEED. So, development and manpower seem likely to explode. Controlling the results to produce a reliable, distributable system that is useful to everyone should not be too difficult.

The Initial Deployment Stage

So, what is going to happen during "initial deployment", and how will we know when that stage completes? Here is how we see it now:

1. We will put up a public server in March. We plan on cross-linking with all of the larger integrations (and as many of the useful smaller efforts as we can). The original WIT effort at Argonne National Lab had to be brought down due to issues that arose in a hacker attack, and we plan on using this initial public SEED as a replacement to which other projects will link. This requires defining a protocol for exposing links. This should be completed in March.
2. We will finally offer versions to any sequencing/annotation efforts that wish to use it (with the understanding that there will be some initial learning and help required to get things going smoothly). The ANL/FIG team will do its best to help this deployment, but we want everyone to realize that this effort is largely a volunteer one. We welcome the experience, since exposing shortcomings and correcting them is exactly what is needed at this stage, but we also expect users to view themselves as participating in the development.
3. We will begin the Project to Annotate a 1000 Genomes.
4. A number of individuals cooperating within the SEED effort will together construct a web service to provide gene calls for prokaryotic sequence. We believe that these will be relatively accurate. The service is planned to reside at the University of Bielefeld in Germany. A second server to provide computation of similarities for genomes to be added to the SEED is planned at Argonne National Lab. Along with the SEED itself, these servers will make it possible for university sequencing/annotation projects to have access to state-of-the-art annotation tools.
5. We will call an end to "initial deployment" when we have
   • successully installed five systems in a row by having users just following instructions (i.e., without our help),
   • have sequencing/annotation efforts that can routinely add new genomes and update versions of old ones,
   • have over 20 users that synchronize weekly using p2p operations, and
   • have a set of at least 10 annotated systems that are routinely distributed and updated via p2p operations.

The Project to Annotate 1000 Genomes

There are differences of opinion concerning what makes the SEED Project important. Ross' view is that the SEED should be thought of as a workbench for producing "subsystem annotations", and that these annotations will eventually be understood to be the most important development growing out of the SEED Project. The roles of

1. supporting initial annotations and
2. helping individual researchers explore genomic data

are important, but not nearly as important as the role of supporting subsystem annotations. At this point, we estimate that roughly 50% of genes in the public archives have solid function assignments, about 20% are completely uncharacterized, and about 30% have either very broad class characterizations or are over specified. Many of the genes within this last 30% have been given accurate characterizations in review articles, but the contents from these review articles often fail to reach the public sequence archives. We propose to use the SEED as a framework for supporting development of "subsystem annotations", which can be viewed as the organized data to be included in a review article. From this perspective, reviews form the essential cutting edge for annotations, and that the SEED should become a vehicle for reducing the effort to produce a review. By providing this service, along with the capability to easily exchange and export subsystem annotations, the SEED will facilitate the flow of assignments from reviews to sequence annotations. A researcher basing his career on analysis of a specific subsystem will have a framework for producing a sequence of reviews, using tools that support maintenance of the subsystem annotation (largely automating the addition of new genomes as they become available).

We believe that there are many, many individuals with extensive experience in a given subsystem that would be willing to produce and maintain one of these "subsystem annotations". Each such curated subsystem would amount to the raw data standing behind a detailed review article. The key issues that must be addressed are roughly these:

• Once a detailed subsystem has been carefully constructed, it must not be lost. This is the most common worry. An expert using a system like the SEED is always concerned that some shift of IDs, new release or whatever will result in loss of the expert's work. By making it straightforward to export an annotated subsystem and exchange them via p2p operations, we believe that we have addressed this issue.
• The system must significantly reduce the effort required to extract and relate relevant data. It is not unusual for an expert to spend years in developing a detailed picture of a subsystem; we believe that this can be dramatically reduced by the development of appropriate tools, but it is essential that the individuals developing tools participate closely in the annotation process; otherwise, it is likely that powerful tools will be developed that do not address the rate-limiting operations.

A minimal notion of "curated subsystem" would include

1. a list of the functional roles included in the subsystem (for metabolic subsystems, this amounts to a list of catalytic domains) and
2. a spreadsheet with genomes along one axis, and functional roles along the other. Each cell would contain a list of the genes in the given organism that implement a specfific functional role.

The rows of the spreadsheet each represent the genes implementing the set of funjctional roles in a given organism, while each column presents a reliable set of genes implementing a specific functional role.

An extended notion would include a number of other items as well:

3. a diagram representing the relationships between the functional roles (in the case of a pathway, a depiction of the reactions that make up the pathway),
4. a discussion of the "variants" represented by the entries in the spreadsheet, and
5. a detailed commentary each set of genes implementing a functional role (describing what can be inferred about the evolutionary origins of the set of genes).

Recently a major effort has been launched to include within the SEED the capability of curating these subsystem annotations and exchanging them via p2p operations.

In March, Ross, Andrei, Veronika and Gary Olsen will begin curation of specific subsystems, synchronizing all assignments and annotations on a weekly basis. Once this initial effort is running smoothly, we intend to rapidly expand the set of individuals participating.

So, in the next newsletter, expect a detailed discussion of the gene-calling and similarity servers, along with a discussion of the outcome of the initial efforts to begin annotation of subsystems.

Finally, our sincere thanks to Dusko Ehrlich and Barny Whitman for sending statements supporting the utility of the SEED. We owe you.

]]> 2006-12-10T13:36:36Z 2004-02-13T18:08:27Z tag:www.figresearch.com,2004:/news//3.110 2004-02-13T18:08:27Z
Meeting in San Diego A Separate Matter...
Veronika Vonstein veronika@thefig.info Newsletters
9. Meeting in San Diego
10. A Separate Matter
]]> We are writing this short newsletter to make some comments on the upcoming SEED Developers Meeting in San Diego (Feb 22-24) and to cover a small number of miscellaneous items.
For those coming to the meeting, it is critical that you establish contact with Andrei Osterman or Ross Overbeek. Their email addresses are as follows:

Ross Overbeek		Andrei Osterman
Ross@TheFIG.info		Andrei@TheFIG.info

The meeting will start at the North Campus of Burnham Institute at 10am on Sunday, Feb 24. The address is

The Burnham Institute

10901 North Torrey Pines Road

La Jolla, California, 92037

858-646-3100

However, we strongly recommend that you link up a bit before that. There is nothing worse than being somewhat disoriented in a strange town. The recommended list of hotels is as follows:

Best Western Stratford Inn

710 Camino Del Mar

Del Mar, California 92014

Tel.: 858-755-1501, x.121 (Dana Hill, Director of

Sales

Fax: 858-755-4704

Toll: 1-800-446-7229

Web: www.pacificahost.com

E-mail: dana@pacificahost.com

Rate: $79 up

Del Mar Inn

720 Camino Del Mar

Del Mar, California 92037

Tel.: 858-755-9765

Fax: 858-792-8196

Toll: 1-800-451-4515

Web: www.delmarinn.com

Rate: $99 up

Hampton Inn

11920 El Camino Real

San Diego, California 92130

Tel.: 858-792-5557 (Mike Macklosky, Sales Manager)

Fax: 858-792-7263

Web: www.hamptoninndelmar.com

Rate: $94

There is a free shuttle service to the Burnham Institute from these hotels, which is what makes them particularly attractive (the prices are pretty good, as well). Ross is planning on staying at the Best Western Stratford Inn and suggests that anyone who wishes meet him for breakfast on Sunday morning at 8am.

This will be a fairly unstructured meeting. The basic rules will be as follows:

1. Participants should come with specific objectives (things they want done or data they want to walk away with). It is their responsibility to formulate a strategy to meet those objectives. For example, Ross has roughly these goals:
   1. He will try to set up a collaboration to call genes (including starts) accurately for prokaryotes. The initial emphasis will probably be for organisms in which some objective (e.g., proteomic) data exists.
   2. The SEED is weak on DNA-related issues. The team at Bielefeld University has built an open source annotation system that is quite strong in these areas. A major topic will be the issue of what it would take to merge systems, exchange tools, or whatever. No serious discussions have yet taken place, but we hope to have some at the meeting.
   3. Ross will arrive with the pre-release version of the SEED that will be modified for an official release about mid March. He will offer copies to those who wish them.
   4. He will demonstrate the new annotation tools/strategy and wishes to seriously discuss how to launch the major annotation effort in March.
   These goals may not match very closely with what other participants want. That is ok.
11. We will occasionally have large discussions covering topics of general interest, but it is expected that smaller groups focusing on specific issues will emerge.
12. If there is a key topic that you wish to speak on (for, say, up to 30 minutes), feel free to prepare a talk. However, we will not be scheduling talks, and our view is that they should happen spontaneously and may be for limited groups.
13. For those people with wireless cards (airports), if you plan on using them at the Burnham, you must send your ID to Andrei Osterman. On a Mac, you can get this from the System preferences/Network settings display. This is critical if you plan to use the airport on Sunday. Andrei needs to get his system support people to help set this up the week before (i.e., next week).

Ross has tried this unstructured approach in the past, and it has worked well. Some of the rest of us are a bit nervous, but we shall see.

A Separate Matter

FIG was recently asked (by a granting agency) to justify efforts to build integrations. The precise request was as follows:

Please describe the usefulness of KEGG, WIT, ECOCYC, and ERGO to the microbiology community

We have already written about what advantages such systems provide, but we believe that the thrust was more "what have they actually done to help the community?" To answer this, we need microbiologists who have actually benefited in some way from one or more of these systems. If any such person would like to write 2-3 sentences, we would be grateful. Please note that we are not asking for comparisons of the systems, just general support for the view that integrations to support comparative analysis will become increasingly significant.

]]> 2006-12-10T13:36:36Z 2004-01-13T19:52:25Z tag:www.figresearch.com,2004:/news//3.122 2004-01-13T19:52:25Z
SEED Tech-Developers Meeting...
Veronika Vonstein veronika@thefig.info Newsletters
14. SEED Tech-Developers Meeting
]]> This newsletter is coming out after only a short break, since we wanted to notify people about the first planned SEED Tech-Developers Meeting. It will be held in San Diego at the Burnham Institute on February 22-24. We realize that this is terribly short notice, but plans were not final until today.

The meeting will be for people actually writing code for the SEED; we intend to hold these meetings to help finalize what goes into the next release. The topics now on the agenda are as follows:

1. We have made a beta release (0.0.0) available to a few people, and comments/criticisms are coming in. These will be reviewed.
2. It has become essential that we immediately address the issue of adding genomes to the SEED, since we now have 20-40 that need to be added as soon as possible. This will include
1. construction of a server to call genes in prokaryotes,
2. reformatting data from several key sources (i.e., building tools to simplify additions from these major sources),
3. how much effort to put into initial automated annotations (and what is wrong with the existing tool), and
4. dividing up the work and responsibilities.
3. We need to finalize the additions to the SEED needed to support the annotation of subsystems. Ross wants to launch the subsystem annotation effort in March, so it is critical that these tools be finalized and blessed at the meeting (they will be reviedwed and corrected in response to criticisms from the annotators, but we need an initial tool that is basically adequate to get things moving).
4. We will begin discussions on the following two topics (the actual coding/integration will begin after the Feb meeting and, hopefully, will move to center stage for the April or June meetings):
1. The SEED does not handle processing DNA well. We need facilities for displaying/aligning/editing regions of DNA. Several users have pointed towards capbilities offered by Artemis as an example. Bielefeld's system has addressed these issues. We need to talk about what capabilities are needed and how to best introduce them.
2. It is clear that the SEED will be used heavily in interpretation of expression data (ESTs, microarrays, and proteomic data). The plans on how this topic will be handled will be talked about over beers.

Other topics can be added, but these should consume most of the three days. We recommend that those who come spend all three full days, but it is not essential. Let us know what your plans/restrictions are, and we will try to work things out. Again, this is a working meeting for groups/individuals who intend to actively participate in SEED development. We are going to schedule another series of meetings targeted towards SEED users and annotators. These should start by March; we will keep you posted.

If you have questions or need help, please contact Ross, Veronika, or Andrei. ]]> 2006-12-10T13:36:36Z 2004-01-01T20:37:30Z tag:www.figresearch.com,2004:/news//3.86 2004-01-01T20:37:30Z

The Project to Annotate the First 1000 Sequenced Genomes, Develop Detailed Metabolic Reconstructions, and Construct the Corresponding Stoichiometric Matrices...
Ross Overbeek ross@thefig.info General The Project to Annotate the First 1000 Sequenced Genomes, Develop Detailed Metabolic Reconstructions, and Construct the Corresponding Stoichiometric Matrices ]]> Introduction

In December, 2003 The Fellowship for Interpretation of Genomes (FIG) initiated The Project to Annotate 1000 Genomes. The explicit goal was to develop a technology for more accurate, high-volume annotation of genomes and to use this technology to provide superior annotations for the first 1000 sequenced genomes. Members of FIG were convinced that the current approaches for high-throughput annotation, based on protein families and automated pipelines that processed genomes sequentially, would ultimately fail to produce annotations of the desired accuracy. We believe that the key to development of high-throughput annotation technology is to have experts annotate single subsystems over the complete collection of genomes. The existing annotation approaches, in which teams analyze a whole genome at a time, ensure that annotators have no special expertise relating to the vast majority of genes they annotate. By having individuals annotate single subsystems over a large collection of genomes, we allow individuals with expertise in specific pathways (or, more generally, subsystems) to perform their task with relatively high accuracy.

The early stages of the effort began at FIG, but quickly spread to a number of cooperating institutions, most notably Argonne National Lab. During the first year of the project, we have developed detailed encodings of subsystems that include a majority of the genes from subsystems that make up the core cellular machinery. More importantly, we have developed the initial versions of technology needed to support the project.

The Project to Annotate 1000 Genomes has reached the stage where it is clear that it will very shortly produce what we call informal metabolic reconstructions that cover the majority of central metabolism as it is implemented in the close to 300 more-or-less complete genomes that are now available. We think of an informal metabolic reconstruction as a partitioning of the cellular machinery into subsystems, the specification of the functional roles that make up each subsystem, and the inventory of which genes in a specific organism implement the functional roles. What is needed to support both qualitative analysis and effective quantitative modeling is to convert these informal metabolic reconstructions into formal metabolic reconstructions. By a formal reconstruction, we mean an accurate encoding of the metabolic network. The goal of such an encoding is to construct a list of metabolites and a detailed reaction network that is internally consistent (in the sense that metabolites that are produced by reactions are connected as substrates to other reactions or to specific transporters, and that all metabolites that act as substrates are produced by other reactions or provided by transporters). Perhaps, a better way to put this is that all apparent anomalies are highlighted as such, and the essential components of the metabolic network are accurately encoded. The output of such an effort is normally what is termed a stoichiometric matrix, the basic resource required to support stoichiometric modeling. One of the central goals of this enlarged effort is to develop accurate stoichiometric matrices for each of the 1000 genomes; we refer to this component of the effort as The Project to Produce 1000 Stoichiometric Matrices.

It is our belief that the development of the technology required to mass-produce accurate genome annotations will ultimately allow fully automated annotation pipelines to achieve relatively high accuracy. Similarly, the existence of 1000 accurate formal metabolic reconstructions would constitute a resource that would allow rapid and accurate development of stoichiometric matrices for newly-sequenced genomes. That is, besides producing accurate annotations, informal metabolic reconstructions, formal metabolic reconstructions, and stoichiometric matrices for a large collection of diverse genomes, we believe that the expanded project will produce technology that will support nearly automatic, very rapid characterization of new genomes.

All of the encoded subsystems, metabolic reconstructions and stoichiometric matrices will be made freely available on open web sites. In addition, the software environments used to develop the encoded subsystems and stoichiometric matrices will be developed and supported as open source software. By making the fundamental data items, the encoded subsystems and stoichiometric matrices, freely available to the community, we expect to stimulate development of alternative software systems to support curation and maintenance of these items.

The Project to Annotate 1000 Genomes

We have chosen to conceptually break the Project to Annotate 1000 Genomes into three stages. We discuss these stages as if they occurred sequentially; in fact, all three stages are now in progress. To understand the three stages, the reader must have at least a rudimentary grasp of what we mean by an encoded subsystem and an informal metabolic reconstruction. When we speak of a subsystem, we think of a set of related functional roles. In a specific organism, a set of genes implement these roles, and we think of those genes as constituting the subsystem in that organism. That is, we are really dealing with an abstract notion of subsystem (in which the subsystem is a set of functional roles) and instances of the subsystem in a specific organism (in which a set of genes implements the abstract functional roles). Precisely the same subsystem and functional roles exist in distinct organisms, although obviously the genes are unique to each organism.

Subsystems are thought of as possibly having multiple variants. Organisms that have operational versions of a subsystem may well have genes that implement slightly different subsets of the functional roles that make up the subsystem. Each subset of functional roles that exists in at least one organism with an operational version of the subsystem constitutes an operational variant.

We think of an informal metabolic reconstruction for an organism as a set of operational variants of subsystems that are believed to exist for the organism. In this conceptualization, one does not have a meaningful functional hierarchy or DAG; rather, we simply have an inventory of functional roles that are implemented in the organism, along with the variants of subsystems that they implement. We do believe that the task of imposing an actual hierarchy is relatively straightforward in comparison with the effort required to construct the set of operational variants. In some contexts, we have included a functional overview in which the subsystems are embedded at the lowest levels. It is clear that, given a diverse collection of informal metabolic reconstructions, the development of appropriate functional hierarchies can be generated with relatively few resources.

Our encoding of a subsystem can now be reduced to

1. a specification of a set of functional roles (this amounts to the abstract subsystem) and
2. sets of genes which implement the operational variants in a number of genomes. These genes are given as a subsystem spreadsheet in which each row corresponds to a single genome, each column corresponds to a single functional role, and each cell contains the set of genes in that genome that are believed to implement the given functional role.

The Project to Annotate 1000 Genomes amounts to an effort to produce detailed and comprehensive encodings of several hundred subsystems, which will impose assigned functions on genes in each of the genomes. The total percent of genes that can be assigned functions this way is probably on the order of 50-70% in most genomes (in large eukaryotic genomes the total is obviously substantially lower). The percent will grow as our understanding grows. What should be noted is that the accuracy of these assignments will be substantially better than that of current assignments, and the conserved cellular machinery almost all falls within the projected subsystems.

Once we have produced our initial set of annotations, we believe that automated pipelines and protein families are excellent tools for propagating them. Protein families are, in fact, a key component of annotation and provide the fundamental mechanism for projection of function between genes. The added dimension provided by subsystems, along with the manual curation required to develop accurate initial encodings of subsystems, is an essential technology for increasing the accuracy and effectiveness of protein families. Ultimately the encoded subsystems will be used to make incremental, essential corrections to collections of protein families (like those supported by UniProt and COGs), and a basis for much more accurate annotation will emerge.

We now proceed to describe the details of the three stages.

Stage 1: Development of Initial Encodings of Subsystems

The initial stage of the project will involve development of approximately 100-150 subsystems that will cover most of the conserved cellular machinery in prokaryotes (and all of the central metabolic machinery in eukaryotes). This work will be done largely by trained annotators who achieve a limited mastery of specific subsystems via review articles and detailed analysis of the collection of genomes. These individuals can define the abstract subsystems and add most genomes to the emerging spreadsheets, but not without error. They are necessarily far less skilled than experts who have invested tens of years in study of specific subsystems.

These initial subsystems will have many uses. They can be used to enhance sets of curated protein families, to clarify identification of gene starts, and to develop a consistent set of annotations. They will form the basis of informal metabolic reconstructions, and will be used to support the development of formal metabolic reconstructions. However, given the relative lack of expertise of these initial annotators and the fact that they will seldom have access to the wet lab facilities needed to remove ambiguities in assignments, errors will inevitably remain.

Stage 2: The Use of True Experts and the Wet Lab to Refine the Encodings

The second stage will involve the gradual refinement and enhancement of the original subsystem encodings by domain experts. Almost every subsystem spreadsheet makes it clear that numerous detailed questions remain to be answered. These questions relate to correcting gene calls, correction of frameshifts, refining function assignments, and removing ambiguities (either via bioinformatics based analysis or through actual wet lab efforts).

The participation of domain experts will be critical, but it seems most likely that a relatively small set will choose to get involved until the utility of the approach becomes obvious. We already have some domain experts (in translation, transcription, and a limited number of metabolic subsystems) participating in the effort. We believe that this number will grow rapidly over the next 2-3 years.

It should be emphasized that upon completion of step 2 we will have accurate annotations and a solid foundation for the construction of stoichiometric matrices.

Stage 3: Understanding the Evolutionary History of the Genes within the Subsystem

The third stage involves determination of the evolutionary history of the genes within the subsystem. To understand what this involves and the utility of this type of analysis, we must simply recommend two papers by the team led by Roy Jensen:

1. Ancient origin of the tryptophan operon and the dynamics of evolutionary change by Xie, Keyhani, Bonner, Jensen, Microbiol Mol Biol Rev. 2003 Sep;67(3):303-42
2. Inter-genomic displacement via lateral transfer of bacterial trp operons in an overall context of vertical genealogy, by Xie, Song, Keyhani, Bonner, Jensen, BMC Biology, 2004, 2:15

These papers elegantly display the exact style of analysis required to uncover and clarify the evolutionary history of the relevant genes. Essentially, trees must be built containing all of the genes implementing each specific functional role (multiple trees may be needed for distinct forms). Those trees that display a common topology indicate which columns in the spreadsheet can be used to infer the most probable vertical history of the subsystem. Once the overall history has been clarified, it becomes possible to attempt clarification of horizontal transfers, to reconstruct the history of clusters on the chromosome, and in some cases to tie the analysis to regulatory issues.

The effort required to do this style of analysis well is high. While we expect the initial efforts to go slowly, we also expect experience and advances in tools to dramatically reduce the required effort. In any event, it is clear that this stage will not be completed in the next few years, but will undoubtedly stimulate large amounts of related research.

Filling in the Missing Pieces

The encoded subsystems produced by the Project to Annotate 1000 Genomes offer a detailed picture of exactly what components have been identified and are present in each genome. Perhaps as significant, they vividly display exactly what is missing or ambiguous, allowing one to arrive at an accurate inventory of gaps in our understanding.

The issue of how best to address these gaps is an integral part of the project. The technology that is emerging is what we refer to as the bioinformatics-driven wet lab. This concept refers to the development of a wet lab that utilizes conventional biochemical and genetic techniques in a framework designed to maximize the overall number of confirmations. It is driven by predictions arising from the analysis of subsystems, and it targets a prioritized list of conjectures. That is, the explicit goal is to fill in as many gaps and remove as many ambiguities as possible for resources consumed.

Although it is inconceivable that one experimental group would be able to assess all of the functional predictions, we believe that integrating an experimental component into our annotation/modeling effort will directly support our main goal. In addition to verification of key predictions and removal of central ambiguities, it will validate the overall approach and set an example for other groups worldwide.

The Project to Develop 1000 Stoichiometric Matrices

We believe that the informal metabolic reconstructions are of substantial value by themselves. Indeed, numerous applications are quite obvious. However, they are not enough to support quantitative modeling. Whole genome modeling will require development of stoichiometric matrices, an effort that will pay many dividends. The most immediate payout is as quality control on the informal metabolic reconstruction. Just as the use of subsystems imposes a critical set of consistency checks on the assignment of function to genes, an attempt to develop an internally consistent reaction network imposes a strong consistency check on both the annotations and assertions of the presence of specific subsystems.

Over the last 4-5 years, the success of stoichiometric modeling has set the stage for large-scale employment of the technology. The key limiting factor is the development of the stoichiometric matrix itself. This is a time-consuming task that frequently requires on the order of a year for a skilled practitioner. Many actual modeling efforts have foundered on just the technical difficulties in producing this basic datum. Bernhard Palsson has pioneered much of the key research that has led to the recent successes. Spending large amounts of effort, his team has built a very few of these stoichiometric matrices, iteratively improving their accuracy. They have successfully used these matrices to support initial modeling efforts on the organisms, and the results have gained international recognition.

PalssonÕs team originated the The Project to Produce 1000 Stoichiometric Matrices, and they will play the lead role in converting the informal metabolic reconstructions into formal reconstructions and produce the matrices. The team at FIG and Argonne National Laboratory will participate in the effort, coordinating closely with PalssonÕs team. At this point, the Palsson team and the teams at FIG, ANL, and The Burnham Institute are all working on issues relating to tools to automate the generation of matrices from informal metabolic reconstructions.

The Participants

We expect participants in both projects from many institutions worldwide, probably with both academic and commercial interests. Initially, it is likely that the effort will be led from FIG, ANL and PalssonÕs team at UCSD. We are planning on Roy Jensen playing a role relating to quality control and development of tools to support Stage 3 analysis. Andrei Osterman from the Burnham Institute will lead wet lab efforts to challenge in silico predictions.

If the effort is successful, we would hope to stimulate numerous research efforts worldwide, and we welcome broad participation. Ultimately, leadership and participation will broaden rapidly, if the effort is successful.

A Proposed Schedule

Let us begin by estimating the point at which 1000 genomes will become available. One simple approach would go as follows:

1. The number of genomes will double approximately every 18 months.
2. We now have about 300 more-or-less complete genomes.
3. Therefore, we should have approximately 1000 genomes in just a bit under 3 years (by sometime in 2007)

There is a great deal in this analysis that is far from certain. However, let us use this estimate as a working hypothesis.

2005

During 2005, Stage 1 will be completed for the vast majority of subsystems. Stage 2 will be initiated for 30-50 subsystems. Less than 10 will move deeply into stage 3.

We will actively attempt to produce 10-15 stoichiometric matrices. We will focus on diverse organisms of interest to DOE and a set of gram-positive pathogens.

We will begin a detailed review for quality assurance by a small number of expert biochemists and microbiologists.

We expect wet lab confirmations to begin, but this is one area in which funding plays an essential role. We expect funding to support targeted confirmation/rejection of the numerous conjectures arising from the bioinformatics to begin in 2005-2006. It is possible to fairly accurately predict the potential flow of confirmations, but we cannot predict available funding. We believe that the bioinformatics-driven wet lab, in which conjectures are prioritized and grouped, would allow a relatively small group (of 3-4 postdocs and technician) to characterize up to 50 novel gene families encoding the most important functional roles in central metabolic subsystems of diverse organisms per year.

2006

During 2006, the vast majority of subsystems will enter Stage 2. We will attempt to move a large number into Stage 3 (this is truly difficult to predict; it depends hugely on success with the early attempts, our ability to reduce the required effort, and the research aims of the participants).

We would plan on completing at least 200 more stoichiometric matrices.

If the wet lab component of the effort is fully functional, we would expect a steady stream of confirmations, and (based on our past experience) we would project roughly that 75-90% of the tested conjectures will be validated.

2007

During 2007 we would plan on pushing Stage 2 and 3 analysis as far as possible. We believe that we will have the subsystems needed to cover the vast majority of well understood subsystems and many that are not well understood.

We would plan on completing initial stoichiometric matrices for several hundred more genomes. Since the majority of the genomes will not become available until this year, of necessity many of the stoichiometric matrices will not be reasonably complete before sometime in 2008 or 2009.

If the wet lab component of the effort is fully functional, we would expect the stream of successful conjectures to stimulate numerous labs to join the effort. Ultimately, the role of the wet lab component that is tightly-coupled to the project is to demonstrate the huge improvement in efficiency that can be attained by coupling the wet lab effort to well-chosen, targeted conjectures generated from the subsystems.

A Short Note on the Analysis of Environmental Samples

It is becoming clear that analysis of environmental samples will become increasingly significant. Consider a framework in which we have 1000 genomes and detailed informal metabolic reconstructions for all of them. We believe that, given a substantial environmental sample,

1. it will be possible to produce accurate estimates of which organisms are present (where an "organism" in this context should probably be viewed as "some organism within a very constrained phylogenetic neighborhood"),
2. it will be possible to produce fairly precise estimates of the metabolism of the organisms believed to be present, and
4. it will be possible to compared the predicted metabolism with the actual enzymes detected in the environmental sample.

The hope is clearly that we will be able to make accurate estimates, given 1000 well-annotated genomes.

Summary

The value of a collection of 1000 genomes depends directly on the quality of the annotations, the corresponding metabolic reconstructions, and the extent to which the foundations of modeling have been established.

The Project to Annotate 1000 Genomes is based directly on the notion of building a collection of carefully created and curated subsystems. The fact that the individuals who encode these subsystems annotate the same subsystem over a broad collection of genomes allows them to gain an understanding of detailed variation and at least a minimal grasp of the review literature. They will be annotating genes for which they develop some detailed familiarity. We place this technology in direct opposition to the existing approaches in which individuals annotate complete genomes (assuring an almost complete lack of familiarity with the majority of genes being annotated), and automated pipelines are badly limited by the ambiguities and errors in existing annotations.

The Project to Produce 1000 Stoichiometric Matrices has the potential of laying the foundations for quantitative modeling. Many, if not most, existing modeling efforts are dramatically hampered by the fact that very, very few stoichiometric matrices now exist, and the cost of developing more using existing approaches is quite high.

The development of a wet lab component that challenges a carefully prioritized set of conjectures flowing from both the subsystems analysis and the initial modeling based on quantitative modeling is essential. It will confirm the relative efficiency of this approach (which might reasonably be characterized as "picking the low-hanging fruit"), and in the process establish a paradigm that directly challenges the more common approach to establishing priorities.

We claim to understand the key technology needed to develop high-throughput development of annotations, metabolic reconstructions, and stoichiometric matrices. By the summer of 2005, this should be completely obvious.

]]> 2006-12-10T13:36:36Z 2003-12-20T17:15:31Z tag:www.figresearch.com,2003:/news//3.106 2003-12-20T17:15:31Z
General Comments on the Schedule for Delivering the SEED An Increase in Programming Support Peer-to-Peer Updating FIG Development Meetings The "Annotate a Thousand Genomes" Efffort...
Veronika Vonstein veronika@thefig.info Newsletters
15. General Comments on the Schedule for Delivering the SEED
16. An Increase in Programming Support
17. Peer-to-Peer Updating
18. FIG Development Meetings
19. The "Annotate a Thousand Genomes" Efffort
]]> General Comments on the Schedule for Delivering the SEED

We promised timely reports on progress, so we thought that it would be a good idea to send out one last newsletter this year. Progress has been very rapid. The SEED was shown at the SC 2003 supercomputing conference, and everything went amazingly smoothly. Early versions of the peer-to-peer updating capability were shown, and a genome was added in real-time. That is, we took a genome with just gene calls, added it to the SEED, computed similarities, made automated assignments, and displayed the results during the demo period. Many thanks to the brave souls that made it happen -- there were points where we wondered whether it could all be made solid quickly enough.

The alpha-test version of the SEED has been installed and operational at several sites for over a month now. We are learning a lot about its utility and what is needed to improve it. Natalia Maltsev is hosting a class on Jan 8-9 that will include two parts: it will present the tools developed by her team to support high-throughput annotation of genomes using grid-technology, and it will present the initial release of the SEED. It will be a small meeting to help make sure that a few sequencing/annotation projects that intend to use the Argonne/FIG annotation tools see what is involved.

We are now working on making a set of DVDs that contain both Mac and Linux versions of the SEED with an install script. We have experienced some difficulties in getting installations to go smoothly, and it is critical that we get the installation script working reliably before a large-scale distribution occurs. In addition, we are finishing the implementation of protocols for sharing annotations. This must all be done and ready to hand out at Natalia's workshop in January.

For those initial users of Mac OS X versions of the SEED, a number of minor problems arose when people converted from Jaguar to Panther. We believe that we understand exactly how to help people make the transition, but be aware that inital versions of the SEED may well stop functioning if you make the switch. Needless to say, we have learned a number of lessons from helping people make the switch, and the versions that will be released in January should run fine under either Jaguar or Panther.

Later in January a class in looking for missing genes will be taught at Franklin and Marshall College in Pennsylvania.

An Increase in Programming Support

Most of the SEED development was done by Ross Overbeek with help from a number of friends. This month Terry Disz and Bob Olson from Argonne National Lab began putting in substantantial amounts of effort, and next month FIG will hire one full-time and one part-time programmer. Part of this expansion is due to FIG receiving its first grant (from DOE). This is a big moment, and we anticipate that it will be just the first of several grants we receive this year.

This represents a huge increase in our ability to support and accelerate development of the SEED. Even so, it is quite likely that a majority of the major advances will come from collaborators over the coming year. This means that you can expect progress to rapidly accelerate (and, really, it has not been so slow up to now!).

Peer-to-Peer Updating

The concept of developing a system that supports peer-to-peer updating is really pretty interesting. The standard way of constructing and deploying systems is based on a central source of both the initial system and updates. The peer-to-peer model is one in which any two users can share and update either data or code.

The classical system lends itself to a hierarchical structure: the main source supplies systems to, say, sequencing/annotation projects. These projects install a single local system that collects and manages function assignments and annotations. Periodically, the project site synchronizes annotations with the main site, propagating work up and then back down through the hierarchy.

In the peer-to-peer model, everyone has a local version of the system, and everyone has the ability to exchange/synchronize with anyone else. Informal "hubs" may form at nodes attempting to maintain current, comprehensive collections of genomes, but everyone has and maintains their own research environment. The way this would work for the average sequencing project would be

20. someone would download (or take from DVDs) an initial SEED,
21. the local genomes would be loaded,
22. other memebers of the project would download copies from the initial version (over a network, through DVDs, or by simply copying between hard drives),
23. periodically (either daily or weekly), members of the project would exchange annotations and assignments.
24. as updated versions of the SEED become available, the new code propagates in a peer-to-peer manner, and,
25. new genomes are integrated and exchanged.

Such an "anarchistic" model introduces numerous questions and potential conflicts, but it does completely remove dependencies on central sources; if you want a new genome immediately, you just add it to your copy (or get it from someone who has already added it). If you have 200 kb of new sequence, you are not faced with sending it to a central source and waiting for them to create an updated version that is returned to you; rather, you just add the data to your copy, do your analysis, and exchange it (or not) with whomever you wish.

This is basically the model we are proposing for the SEED, and the initial "launch" is very close.

FIG Development Meetings

The FIG developers will meet at least three times per year (one week meetings) to integrate code and prepare releases. We anticipate having two meetings per year in the US (one in Chicago) and one per year in Europe. Because there is so much to do initially, we are proposing to meet three times in the next 6-7 months. The first meeting will probably be in the second half of February, the second in Europe in April, and the third in at Chicago in June. The locations of the first and second meetings are still not decided, although the second will probably be in Bielefeld, Germany. The first cannot be in Chicago for two reasons: the weather in February can be terribly unpredictable, and it takes 90 days to get clearances for foreign visitors. We plan on having these meetings be completely informal code integration efforts. People will focus on defining the future additions to the FIG software, integrating software from other independent efforts, and preparing new releases. We anticipate releasing new versions of the FIG software about a week after each meeting. If you have suggestions regarding schedule or location, comments would be welcome.

The "Annotate a Thousand Genomes" Efffort

Developing the SEED is a major effort initiated by FIG, and we believe that it will be of utility for numerous projects. Initially, we believe that it will be used largely by sequencing/annotation teams and individual investigators wishing to use comparative analysis using large collections of genomes. However, a number of FIG fellows wish to focus on the problem of developing reliable annnotations, and they believe that they finally see exactly how to do it. We have discussed this briefly in past newsletters. The project will involve gathering experts in specififc subsystems and supporting analysis of each subsystem accross the entire collection of organisms. We would add whatever functionality is needed to the SEED to support these experts, and the explicit goal would be accurate annotations for the thousand genomes we believe will be available by the end of about three years.

This effort, too, is beginning to make progress. Most notably, a number of biologists with specific areas of expertise have expressed a desire to paricipate actively, as soon as the project is a bit more well-defined, and the tools are in place. We intend to officially launch things in February.

Our first step is to define precisely what "deliverables" we would want from each expert. This may sound like an insignificant issue, yet a number of us are now convinced that getting it right is the key to saving large amounts of effort. We are actively discussing the point, and we will offer an initial position in the next newsletter.

Once the initial release of the SEED occurs (hopefully within a very, very short time now), expect to see focus rapidly shift to defining and starting the annotations effort.

1. We have built a "SEED disk" for Mac OS X machines. We have built six copies so far; debugging and improvements are progressing very quickly.
2. We have gotten our new dual-processor G5 server. It is a wonderful machine. We are using it for development and the SEED runs well on it. We will put up a public server as soon as we load it with completely current copies of all of the 140-150 published genomes.
3. The SEED will be shown at Supercomputing 2003 (a big yearly show in the world of supercomputers) later this month. We will have two G5 Macs and two linux machines (one running Postgres, and one running MySQL). As of today they support peer-to-peer (p2p) updating of code, annotations and assignments.
4. By next week we plan on supporting the ability for users to add their own genomes, and probably the ability to acquire initial copies over the internet (we will not immediately offer this service -- it will take at least a month to get the initial release ready).
5. We have a problem getting data entered initially. We have found at least 2-3 groups who will help us; one by helping acquire and prepare published genomes, one might help us call prokaryotic genes, and a third might help us organize the eukayotic data. I hesitate to announce names, since the commitments are not firm yet (I believe that everyone is waiting for us to organize our side of the collaborations). However, we do deeply appreciate the offers of help, and we believe that the pipelines will be flowing at high volume by the end of the year.
6. The SEED is now actively in use by a large-scale sequencing effort. It will be used as the framework for analysis of genes and development of metabolic overviews. We have been approached by a number of other groups who are now planning on using the system, if their grants go through. These are all important, since each such group will become a valuable source of criticism. There is a real difference between the demands of people who just occasionally use the system and those that use it on a daily basis. The speed of improvement will be directly proportional to the number of serious users.
7. The system is being used to support a class at the University of Chicago. The real impact of this effort was to force a bit more organization in our development. Having a class of students dependent on the use of the system is exposing bugs and shortcomings pretty quickly. FIG is actively planning on helping to develop course materials and to aid in sharing of course materials. We feel that it is important that we build up a collection of web-based "lectures" and "assignments". Initially, we will focus on developing three 2-week "modules" that would be suitable for use in any number of biology courses. The modules will be
• Searching for Missing Genes
• How to Annotate a Genome
• How to Annotate a Subsystem

So, the SEED now runs on Linux and Macs. The next month will be spent in getting installation procedures and instructions simplified, as well as loading the data for the initial release. Once the initial release occurs, we expect a very rapid propagation. If our ability to support peer-to-peer updates of code is solid, we will be able to handle bug fixes and new features without it being a big deal. If not, we will rapidly be overwhelmed. There is, of course, an ongoing tension between making things solid and adding new features; it always exists, and we are trying to make the right choices.

The "annotate a 1000 genomes" project is moving from the discussion stage to becoming a clear goal of FIG. The basic position may be summarized as follows:

1. Think of a spreadsheet with each row corresponding to a genome, each column corresponding to a subsystem (e.g., "synthesis of histidine"), and each cell corresponding to a summary of how the given subsystem is implemented in the specified organism. The number of rows will grow at ever increasing speeds, but the number of columns (i.e., subsystems) is relatively constant.
2. Most current annotation efforts focus on annotating a single genome or a small set of closely related genomes. The key to "high-throughput annotation" is not the pipeline approach that is widely employed, but rather the development of a detailed understanding of the variants of a specific subsystem, followed by consistent annotation of that subsystem through all existing genomes (i.e., annotation of columns, rather than rows).
3. There are on the order of 80-120 subsystems that make up the "core machinery" of the cell. Here, we explictly exclude a few key categories that should be treated separately (e.g., signalling, transport, and regulatory genes). The study of each of these subsystems needs to be done systematically.
4. One of the best examples of how this should be done, in our opinion, was published in the Sept, 2003 issue of mmbr. The paper, "Ancient Origin of the Tryptophan Operon and the Dynamics of Evolutionary Change" by Roy Jensen's team, is a vivid portrait of exactly what is needed for each of the key subsystems. We highly recommend it; it is an excellent review of the tryptophan biosynthesis subsystem, but even more importantly it is a precise template of what is needed for each subsystem.
5. The speed with which an analysis of the sort discussed in the tryptophan review can be constructed is highly dependent on the tools the analyst uses. A good biologist might well also suggest that it depends on insight formed from years of work on a specific subsystem, but it seems clear to us that the availability of appropriate tools is a major issue. The right tools do not yet exist (at least in our view). FIG will partiipate actively in the development of these tools.
6. FIG does not intend to limit its participation to tool building. Indeed, several of us are deeply interested in actively taking subsystems, performing the needed analysis, cleaning up the annotations, and producing the corresponding reviews.

The FIG team is growing rapidly (in the sense that a lot of interesting people are starting to closely collaborate on this strange effort). It is most encouraging.

More in a few weeks.

Take care,

The FIG Team