0
$\begingroup$

I am trying to narrow down the possible search criteria from an RNASeq experiment. I already have the cds sequences so I took these and used Blast+ program to blast the data against itself to find sequences that were the same.

From what I have read the E Value is based on the database size. My database only has ~27,000 sequences in it, does this mean that only exceptionally low E values will be relevant to me? Are there any other criteria I should search for before writing the program to eliminate similar sequences?

$\endgroup$
1
  • $\begingroup$ What are you mapping to the database- RNAseq reads? How long is the query? $\endgroup$
    – WYSIWYG
    Aug 19, 2015 at 5:21

1 Answer 1

0
$\begingroup$

Going by an e-value or bitscore cutoff will give you possible homologs, but it sound like you want to remove redundant sequences. If you want to cluster and combine similar sequences to make a smaller database, you could just go via sequence identity using something like CD-HIT. This, for example, is what's done to produce the UniRef set from the UniProt database.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .