the relationship between query sequence length and e-value/bit score in NCBI BLAST+

Question

[bill-zt]

During a practice, I used 30 DNA query sequences in equal length (100bp). They are all from human genome.

After BLASTing them to the human reference genome, I saw an intersting result: although the sequences are totally different, the best match of each query always has the same e-value and bit score.

The e-values are always 2.67e-45

The bit-score are always 185

Since all the 30 DNA query sequences are completely different except they are in equal length, I guess whether the e-value and bit-score just depend on query sequence lengths? Is there any exact relationship rule or formular?

Answer 1

[Jean-Karim Heriche]

Read about the BLAST heuristics and associated statistics, for example:

• NCBI's The Statistics of Sequence Similarity Scores
• chapter 4 of the book BLAST - an essential guide to the basic local alignment search tool.
• the review Issues in searching molecular sequence databases.