Genome mapping and clustering

 

The 5’- and 3’-end cDNA sequences and the full-length cDNA sequences were mapped onto the human genome*, UCSC hg 17 NCBI Build 35 or UCSC hg 18 NCBI Build 36.1. Possible local alignments between cDNA and genome sequences were detected by NCBI Mega BLAST (1), ftp://ftp.ncbi.nih.gov/blast/. The best mappings of each cDNA sequence were constructed from these local alignments using a dynamic programming technique so that the identity, coverage and topology of exons were optimized. The joints of consecutive local alignments were refined so as to restore consensus sequences in canonical splice sites. Clustering of cDNA sequences were done based on the mapping results: two cDNA sequences were grouped into the same cluster if their mapped positions shared at least one base on the genome. Mostly, each cluster corresponded to a single gene locus.

 

* The human genome used:

    FLJ human cDNA database ver.1.0 & 2.0, UCSC hg 17 NCBI Build 35

FLJ human cDNA database ver.3.0, UCSC hg 18 NCBI Build 36.1

** Identity and coverage:

    Identity, >=95%

    Coverage of consensus length, >=50%

 

References

(1) Zhang, Z., Schwartz, S., Wagner, L., and Miller, W. (2000) J. Comput. Biol., 7: 203-214.