comparing two genomes

Question

[dr_mm_sh]

I am trying to find low similarity DNA sequences from two genomes of very closely related fungal species. Blast+ alignment of concatenated genomic sequences showed hits with identities around 70~80%. However, when I retrieved these DNA sequences of the both query and subject and made blastn search against genomes of each other, I was able to see hits with 100% identities.

What do you recommend to improve the accuracy of alignment results?

Command used to produce the alignment was:

$ blastn -db db_file -query input_file.txt -out output.txt -max_target_seqs 100 -outfmt 7

Answer 1

[edrezen]

Hello,

The blastn command actually executes the megablast algorithm by default. The megablast algorithm is very fast but finds fewer results than the blastn algorithm.

If you want to execute the blastn algorithm, you should try the following :

blastn -task blastn -db db_file -query input_file.txt -out output.txt -max_target_seqs 100 -outfmt 7

This post talks about this "-task" option.

Answer 2

[Sishuo Wang]

discontinuous megablast might also be helpful or you can change '-word_size' (=11 be default).