Need help in BLASTP parameters (dbsize & searchsp)

Question

[shiv]

Hi,

I have a large database for blast analysis so I decided to split the database into smaller chunks and blast the query sequences against them. I am also thinking about to look the difference of E-values between those blast results (query vs small chuck DB and query vs the whole database). After reading the manual and so many discussed points here about the -dbsize parameter in BLAST. I have some questions regarding the parameters :

1: Should I use the -dbsize and -searchsp together at the time of blasting against the small chucks DB file (if i want to use the size of actual whole database)?

2 What should I provide in -dbsize parameter, total number of residues in DB OR total number of sequences in DB ?

3 What type of value is provided in -searchsp parameter ?

Thanks in advance !!

Answer 1

[Dattatray Mongad]

You can refer the for more information on effect of dbsize, searchsp and query length on blast results in SS-Wrapper: a package of wrapper applications for similarity searches on Linux clusters. But, I think you need not have to worry about dbsize parameter as it is used for calculating e-values only. If you want your e-value to be calculated on whole database then keep it as 0.

Answer 2

[Mensur Dlakic]

In my experience the total number of sequences is the most important parameter to set for your case, and this is done with-dbsize.