The Community for Technology Leaders
Green Image
Issue No. 04 - October-December (2009 vol. 6)
ISSN: 1545-5963
pp: 652-666
Hiroaki Uehara , Keio University, Yokohama
Masakazu Jimbo , Nagoya University, Nagoya
The study of gene functions requires high-quality DNA libraries. However, a large number of tests and screenings are necessary for compiling such libraries. We describe an algorithm for extracting as much information as possible from pooling experiments for library screening. Collections of clones are called pools, and a pooling experiment is a group test for detecting all positive clones. The probability of positiveness for each clone is estimated according to the outcomes of the pooling experiments. Clones with high chance of positiveness are subjected to confirmatory testing. In this paper, we introduce a new positive clone detecting algorithm, called the Bayesian network pool result decoder (BNPD). The performance of BNPD is compared, by simulation, with that of the Markov chain pool result decoder (MCPD) proposed by Knill et al. in 1996. Moreover, the combinatorial properties of pooling designs suitable for the proposed algorithm are discussed in conjunction with combinatorial designs and d\hbox{-}{\rm disjunct} matrices. We also show the advantage of utilizing packing designs or BIB designs for the BNPD algorithm.
DNA library screening, high-throughput screening, pooling experiment, two stage test, group testing, Bayesian network, sum-product algorithm, LDPC.

H. Uehara and M. Jimbo, "A Positive Detecting Code and Its Decoding Algorithm for DNA Library Screening," in IEEE/ACM Transactions on Computational Biology and Bioinformatics, vol. 6, no. , pp. 652-666, 2007.
81 ms
(Ver 3.3 (11022016))