67. Construction of a Genome-Wide Human BAC-Unigene Resource 

Bum-chan Park¹, Robert Xuequn Xu, Chang-Su Lim, Mei Wang, Aaron Rosin, Steve Mitchell, Hee Moon Park¹, Eunpyo Moon², Ung-Jin Kim, and Melvin I. Simon 
Division of Biology, Caltech, Pasadena, CA 91125 
¹Chungnam University, Taejon, Korea and ²Ajou University, Suwon, Korea 
simonm@cco.caltech.edu 

With the availability of high quality BAC libraries with stable, large inserts, it is now feasible to rapidly develop genome-wide physical BAC contig resources to cover the large mammalian genomes. For this purpose, we have tried to screen human BAC libraries using mapped Unigene cDNA clones as probes. Currently, over 52,000 mapped Unigenes (non-redundant, unigene sets of cDNA representing EST clusters) are available for human alone. A total of 44,000 Unigene cDNA clones have been supplied to us by Research Genetics. We have currently deconvoluted over 10,000 Unigene probes against a 4X coverage human BAC library D using high density colony hybridization filters. 10,000 batches of Unigenes are arrayed in a logical array of 100 X 100 matrix from which 100 row pools and 100 column pools are derived. Library filters are hybridized with pooled probes, thus reducing the number of hybridization required for addressing the positives for each Unigene from 10,000 to 200. Details on the experimental scheme as well as daily progress report is posted on our WEB site (http://www.tree.caltech.edu). Initial assessment of the deconvolution data indicates that over 95% of the Unigenes have been deconvoluted so that we could have made a BAC-Unigene resource for them. 800 additional Unigene probes and 1,200 Unigene probes which were already deconvoluted by 100x100 have been re-screened by 20x20 to determine the accuracy and to estimate the rate of false positive hits as a function of probe complexity and improve the accuracy. To circumvent the cross-hybridization problems inherent to some Unigene probes, we are also designing OVERGOes from sequences derived from mapped, well annotated genes. Human BAC-Unigene resources generated in this effort will contribute toward the realization of the "whole genome" approaches for human and other model organisms.