Poster: Late and Moved Abstracts

Abs # 1035: Browsing the soybean genome: Educational challenges from physical map builds of a recently duplicated genome

Presenter:	Shultz, Jeff , jlshultz@siu.edu
Authors	Shultz, Jeff  (A) (E)  Meksem, Khalid  (A)   Langin, Chet  (A) (D)  Kazi, Samreen  (A)   Zobrist, Kimberly  (A)   Yaegashi, Satsuki  (A)   Lavu, Nagajyothi  (A)   Iqbal, Javed  (A)   Potter, Jamie  (A)   Yesudas, Charles  (A)   Wainer, David  (F)   Watson, Dennis  (A)   Wu, Chencang  (B)   Zhang, Hong Bin  (B)   Town, Christopher  (C)   Lightfoot, David  (A) (D)
Affiliations:	(A): Genomics Core-Facility, Southern Illinois University at Carbondale, Carbondale, IL 62901-4415 USA (B): Department of Soil and Crop Sciences, Crop Biotechnology Center, Texas A & M University, College Station, TX 77843-2123 (C): The Institute for Genomic Research, Maryland, USA (D): Department of Medical Biochemistry and Molecular Biology, SIUC (E): Department of Plant Biology, SIUC (F): Dept of Computer Sciences SIUC
Web Site:	http://soybeangenome.siu.edu/

Communicating physical map structures for genomes that recently increased ploidy present challenges for scientific community education. Problems arise from the conditional nature of linkage group assignments. The physical map of the soybean genome (build 3) had 388 markers, anchoring only 30% of 2,903 contigs unequivocally. In build 4 anchored markers increased to 404 among high quality fingerprinted BACs. Clones suspected of contamination were listed, examined and contaminates removed from FPC analysis. The edited database produced 2,854 contigs and encompassed 1.05 Tbp. Homeologous regions were detected in two forms. First, contigs with high clone density (23%) were inferred to represent 2-4 conserved genome regions and were labeled for easy recognition by users. Second, pairs of separate contigs that contained the same marker anchors (69%) were inferred to represent homeologous but diverged genomic regions. G-browse was adapted to represent homeologous regions by showing each contig at all potential anchor points, spread laterally and prevented from overlapping. Users have strong visual cues to use caution when interpreting data in these regions. For ease of gene mapping a minimum tiling path was extracted that contained 9,546 clones. The 13,474 BESs from the minimum tiling path showed evidence of gene rich islands on each chromosome. EST probes from 313 low copy genes identify ~2000 BACs. Gene preditions of both types are shown on separate tracks and link to GenBank. QTL link to Soybase. We conclude the representation of physical map builds within recently duplicated genomes can be improved with existing databases and tools but that new algorithms are needed for genome representation in G-browse to a wider community. This work was supported by NSF project #9872635