The purpose of this second Consortium meeting was to continue to coordinate effort among groups committed to generating a publicly available Full-Length cDNA sequence resource. This would build on the success of the I.M.A.G.E. Consortium EST Sequencing Program in avoiding duplication of effort.
This meeting followed a successful meeting held in May following the Cold Spring Harbor Genome Sequencing and Mapping Meeting. [Note: For additional relevant links and background material on complete cDNA sequencing, see the report of the first Workshop on Complete cDNA Sequencing (May 19, 1997).] At the May consortium meeting, the various participants outlined their efforts, which largely consisted of pilot projects. There was universal recognition of the value of the continuing full-insert and full-length cDNA sequencing efforts, but most groups had little, if any funding support for their projects.
Attending this meeting were representatives of both academic and commercial cDNA sequencing groups. Group representatives presented summaries of their work, but these presentations were interspersed with general discussion of a variety of issues, including the distinction between full-length and full-insert sequencing projects, the need for coordination to avoid duplication, and the best manner in which to implement a coordination/registration database.
Funding for the European cDNA sequencing project began September 1, 1997. A total of 400,000 ECUs has been provided to each of 8 institutions. Participating institutions (see EURO-IMAGE below) will generate 8Mb of cDNA sequence spanning the full inserts of cDNA clones. There are no plans at present to create full-length cDNA libraries or to try to identify full-length clones. There are also no universal plans to coordinate sequencing at present, although Charles Auffrey is working with a number of groups to coordinate efforts. (After reading a draft of this report, Auffrey sent a description of the EURO-IMAGE project.)
Michael Rhodes, UK-HGMP Resource Center
Using Washington University's icatools, the UK-HGMP clustered all the IMAGE clone cDNA
sequences and 60,000 clones believed to be unique were identified. These are currently being
rearrayed as a unique collection that will be made publicly available. However, the HGMP does
not presently have funds to sequence these clones. Rather, they will attempt to coordinate with
the European Consortium to have these clones sequenced.
Karl Guegler, Incyte Pharmaceuticals
Incyte is continuing to sequence ESTs as well as full-length clones. They are currently
considering a plan to release 3' and 5' sequence data and issuing public notifications of clones that
are about to be fully sequenced. Most of this will occur through Genome Systems, Inc., which
was recently purchased by Incyte.
David Smoller, Genome Systems, Inc.
Genome Systems has clustered the IMAGE sequences, producing 23,000 clusters and 33,000
singletons. They are now identifying the longest 5' clones from each of the clusters and plan to
rearray approximately 18,000 of these long clones and provide both clones and filters at a
relatively low cost. In addition to the human clones, they are also rearraying Drosophila
clones.
Richard Gibbs, Baylor College of Medicine
Baylor is continuing to sequence full-insert cDNA clones using the concatenation strategy that
they developed. The goal is produce 1000 full-insert cDNA sequences; 150 have already been
completed and submitted to GenBank and another 700 are in the pipeline from IMAGE plates 75
and 76.
LaDeana Hillier and Jeff Woessner,
Washington University
Wash U is attempting to sequence the full inserts from clones representing the UniGene clusters,
but avoiding those clusters for which existing complete mRNA sequence exists. They are
choosing clones with inserts smaller than 800 bp in size so that the clones can be sequenced
completely sequenced using long reads from the clone 3' and 5' ends. At present, they have
identified 8060 and have 400 completed that they are preparing for submission to
GenBank. They will post the IMAGE IDs of the clones they plan to sequence on the Washington
University web site. Wash U has also secured Howard Hughes funding to use the same strategy
to sequence one representative full-insert clone from each unique Mouse cDNA cluster.
Stefan Weimann, DKFZ
As part of the European Consortium project, DKFZ has begun sequencing of 384 clones larger
than 1.5 kb (and an average insert size of approximately 2kb). They have also constructed 4 new
libraries, 3 using the Clonetech CAPFINDER technique and 1 using "standard" cDNA library
construction protocols.
Horst Domedy, Genzentrum, Muenchen
As part of the European cDNA consortium, Genzentrum will generate 15,000 ESTs as well as
1Mb of additional cDNA sequence, focusing on Chromosome 21-specific cDNAs. These will be
make publicly available through the German Genome Resource Center.
Babru Samal, Amgen, Inc.
Amgen is looking at signal peptide-producing genes in mouse, identifying them through EST
sequencing and converting them to full-length sequences using RACE/Clonetech Marathon
protocols. However, there are at present no plans to make either the clones or their sequence
public.
Takao Isogai and Toshio Ota, Helix Research
Institute, Japan
Helix and Dr. Sugano of the University of Tokyo, have constructed libraries with enhanced
representation of full-length clones using an oligo-capping method and full-length sequencing has
begun using these libraries. Neither the libraries nor their sequences are publicly available at this
time, but they are constructing a plan to release both. Dr. Nomura's group at the Kazusa Research
Institute, Japan, is sequencing approximately 2Mb of cDNA sequence per year representing about
300 cDNA clones. Dr. Sugano of Tokyo is also constructing mouse full-length cDNA libraries
and those will be publicly available.
Giorgio Valle, University of Padua, Italy
Padua has done their own sequence clustering. From the unique clusters, they have mapped
approximately 500 clones on the GenBridge 4 Radiation Hybrid Mapping panel, and are
attempting to obtain expression data for these transcripts. They are collaborating with Charles
Auffrey to identify candidate full-length clones using a PCR strategy to screen 20,000 pooled
clones in to identify the longest representative of these clusters.
Robert Cottingham, Genome Database
As IMAGE clone libraries are registered, Greg Lennon submits these to GDB, so all IMAGE
clones have GDB accession numbers (GDBids). In order to cross-reference mapping and
sequence data, GDB is establishing links between dbEST and cDNAs in GDB. This effort would
be greatly simplified if EST sequence submissions to dbEST included the GDBid for the
corresponding cDNA clone in GDB. A suggested submission
protocol was provided.
Levy Ulanovsky, Argonne National Laboratory
Ulanovsky and collaborators at ANL have been applying their DENS Oligowalking approach to
the sequencing of cDNA clones. This technique has promise to reduce the cost of primer walking
sequencing approaches by eliminating the need for custom primers.
John Quackenbush, The Institute for Genomic Research
TIGR has been working on a number of approaches to creating full-length libraries. The 5' CAP
trapping protocol described by Carninci et al. (Genomics 37, 1996) has proven difficult to
reproduce so a number of other approaches are being attempted. In order to begin generating
full-length cDNA sequence, the TIGR Tentative Human Consensus (THC) assemblies are being
used to identify candidate full-length clones. At present, 50 have been identified and completely
sequenced, producing 37 full-length sequences representing 36 unique transcripts. An
additional 13 full-insert sequences have also been completed but represent only partial transcripts.
1. The need for full-length libraries.
2. The need for additional funding for full-length and full-insert sequencing.
3. An understanding of the various clustering methods and their relationships.
4. Coordination of sequencing efforts in order to avoid duplication.
While all present agreed that we should continue to work on the first three items, it was clear that it was appropriate to create an allocation database similar to the RHalloc database at EBI used to coordinate Radiation Hybrid Mapping. The following data were suggested as necessary for each clone registered in such a cDNA allocation database: clone IMAGE ID, 3' and 5' sequence Accession Numbers (extracted from dBEST), the estimated clone length, whether the clone is full-length or simply sequenced for its full insert, the date the clone was registered, and a status. Allocations would continue until the full-insert sequence was submitted to dBEST, at which time the status would change from "allocated" to "finished"; if an allocated clone is not sequenced in a timely manner, the status should change to "expired". There was some discussion about how long a registration should be allowed to remain without the clone being finished before it expires. The consensus was that six months should be sufficient to finish sequencing. The other issue that was not resolved was the number of clones that groups should be allowed to register at any one time. Most felt that 384 clones was a good initial allocation size. Bob Cottingham agreed to contact Particia Rodriguez-Tomé of the EBI to discuss the possibility of establishing such a database.
While still at Hilton Head, Dr. Rodriguez-Tomé contacted me to ask that I let the IMAGE consortium know that she would be happy to establish such a database provided that funding be made available to support it. Dr. Rodriguez-Tomé was preparing a grant application to support a similar genomic sequencing registration database, and she felt that a full-length cDNA sequencing database would be a natural extension of her application. As the application was due immediately following her return to the EBI from Hilton Head, she asked me to provide a letter of support on behalf of the IMAGE Consortium.
Return to Genome-Related Meetings
Please send questions or comments to martinsa@ornl.gov