[Biodevelopers] Re: BLAST asymmetrical

[Michael Nuhn]

Hi, Ela!

>BLAST is a heuristic, so it is not just asymmetrical,

I am still wondering: which part of the algorithm makes blast asymmetrical?

>but can miss hits of interest, if you are not lucky.

This virtually never happens. It is more a question of thresholds used,
which, I admit, is a topic that can lead to lengthy discussions.

I have compared Nucmer and Blast alignments for aligning contigs to a
reference genome and found that Blast produced better results (for us)
because it is more sensitive. (Which is not really a surprise.)

>You could use MUMmer to align two genomes.
>http://mummer.sourceforge.net/.

Mummer is a great program to get an overall overview of rearrangements in a
pair of genomes. I am interested in where genome A and genome B have
homologies and where they don't. The Mummer alignment algorithm discards
hits in the stage where it constructs the longest increasing subsequence.

http://mummer.sourceforge.net/MUMmer.pdf (Section: Sorting the MUMs, Figure
3)

I want all homologies.

>Or even Smith-Waterman which will take a while to run.

Do you know of a program that can calculate SW on a pair of genomes?

Greetings from Germany,
Michael.