70. High Density Colony Filter Production and Automated Data Analysis for Efficient Hybridization Screening of BAC Libraries 

Anca Georgescu, Laura Kegelmeyer, Bernadette Lato, Hummy Badri, Matthew Groza, and Anne Olsen 
Human Genome Center, Biology and Biotechnology Research Program, Lawrence Livermore National Laboratory, Livermore, CA 94551 
olsen2@llnl.gov 

Bacterial Artificial Chromosomes (BACs) have proven to be excellent reagents for construction of sequence-ready maps. As the demand for mapped BACs continues to increase, efficient methods for screening these libraries are needed to identify clones at the required throughput. We describe a format for high-density colony filter production in conjunction with a program for automated analysis of hybridization results, which has greatly improved the accuracy of identifying positive signals. 

Colony filters are plated at a density of 6 x 6 x 384, or 13,824 colonies per 8 x 12 cm filter. All colonies are plated in duplicate in a unique offset pattern optimized to prevent ambiguity in identification of positive offsets. A positive control is plated in the first and last offsets of each subgrid to enable automated drawing of major grid lines by the analysis program. Hybridization probes consist of Alu-PCR products of cosmid or BAC clones, or overgos (J. McPherson, Washington Univ.) designed from cosmid or BAC end sequences. Multiple probes are pooled in a single hybridization, and positive colonies are re-arrayed for hybridization with individual probes. Hybridization signals are analyzed by the "blot-score" program developed at LLNL. The program currently operates in semi-automated manner, with the potential for full automation in the near future. In the current mode, the user loads a phosphorimager file of hybridized filters and selects the filters to be analyzed. Using the positive control offset signals, the program dynamically draws gridlines to indicate the 384 major subdivisions on the filter. When a subgrid containing a positive is selected, the program displays an enlarged view of the positive subgrid with the 6 x 6 minor gridlines drawn. True positives appear in unique duplicate patterns. The program highlights the position of the expected duplicate signal when the user selects a positive, thus providing immediate feedback on the validity of a given pair of observed signals. The program is linked to the LLNL mapping database to facilitate entry of hybridization results into the database and retrieval from the database of map information relevant to positive clones identified. 

Work performed under the auspices of the US DOE by Lawrence Livermore National Laboratory under contract W-7405-ENG-48.