Blast Parameters For Searching Rna Database

Question

[Woa]

I've a collection of cDNAs (median length ~1500 ) and I wish to filter out those are RNA coding ones thru a blastn search. Can people tell me what typical blast parameters are to be chosen for that kind of search? What will be the typical cutoff values for bit score, evalue, alignment length, low complexity masking(default or soft) etc.?

Thanks in advance

[Jarretinha]

Have you considered using Infernal instead of blast?

[Woa]

Thanks and yes,I'm looking into Infernal and Rfam

Answer 1

[Marina Manrique]

One easy thing to do is to pick one RNA from the blast database and search for it in the database, then look at the values of evalue and bitscore of the hit with itself. This can give you an idea of the range of the values for an exact match.

Giving an absolute threshold value for the evalue or the bitscore can be a bit risky since they normally depend on the size of the blast database, the size of the query set...

Answer 2

[Peter]

If you want to search for RNAs with blast, maybe also consider these parameters for the local alignment step: http://stevemount.outfoxing.com/Posting0004.html

[Woa]

Thanks !!!, I've also found another guideline but that's for Prokaryotes( my sequences are mammalian) http://sigen.org/index.php/sigen/article/viewArticle/sigs.651139/192