blast command line and online

Question

[snehalp]

Hello :

I am trying to run the refseq blastp command

> blastp -query short_test.fa -db human_refseq -num_alignments 1  -matrix pam30 -gapopen 5 -gapextend 2  -evalue 10  -outfmt '6 qseqid sallseqid sseq mismatch pident qlen length evalue bitscore qcovs' -out short_refseq.txt

THis is the query I am running and below is the info of my database which in only human :

> Database: sequence.fasta  113,979 sequences; 76,014,371 total residueDate: Oct 31, 2018  3:28 PM  Longest sequence: 35,991 residues

Now my problem is online version gives entirely different result to command line version : Command line finds bitscore of only 43.3 while online version finds 90.6 bitscore and peptide(SELPLQEQNYPAVDSPPSGGGWAGWGSWGK)is matched exactly in online version(SEIPLQEQNYLAVDSPPSGGGWAGWGSWGK ) vs command line (SEIPLQEQNYLAVDSPPS) for the same protein.. for example :

>   command line 
> SELPLQEQNYPAVDSPPSGGGWAGWGSWGK    NP_001337561.1  SEIPLQEQNYLAVDSPPS  2   88.889  30  18  3.85E-06    43.3    60  
> 
> 
>online version  maps to exact peptide : 
  SEIPLQEQNYLAVDSPPSGGGWAGWGSWGK 
> Score Expect  Identities  Positives   Gaps
90.6 bits(206)  3e-22   28/30(93%)  28/30(93%)  0/30(0%)

Can anyone help me with parameters ..T

[GenoMax]

my database which in only human

Are you limiting online search to the same database (human_refseq)? Online blastp uses BLOSUM62 by default. Gap open penalty is 11 and gap extension is 1online. These are all different in your command line search.

[snehalp]

I tried using gap penalty to 11 and 1 too..bitscore reduced to 33 further with that gap value ..Becuase its short sequence online blastp changes the matrix to pam30 ..

[GenoMax]

That is good to know. That leaves the database you are using as possible key difference. Are you using the latest v. 2.8.1?

[snehalp]

mine is 2.7.1..will that be problem ? I updated the version but still same results.