BOL

The key to finding a solution is to notice that most genomicsequences differ by very little. It may well be that the number of complete genome sequences being stored is increasing rapidly, but the actual amount of new data is very small. In other words, a single DNA sequence isn't particularly compressible but a set of sequences shares so much in common that the redundancy can be used to store them in a much smaller storage space. (Source:e-article from Alex Armstrong)

http://www.i-programmer.info/news/181-algorithms/4537-a-new-dna-sequence-search-compressive-genomics.html

http://en.wikipedia.org/wiki/Compression_of_Genomic_Re-Sequencing_Data

http://www.nature.com/nbt/journal/v30/n7/full/nbt.2241.html

http://bioinformatics.oxfordjournals.org/content/29/13/i283.full

http://groups.csail.mit.edu/cb/cast/