+1 In addition to these points there may also be some formatting issues with the input, given the error messages. I think that formatting the db correctly and using the correct db for the program will probably solve this problem though.
Hello, I have a fasta file with 7600 nucleotide sequences, I have a local nr database formated for blast, with makeblastdb, I don't use formatdb because it give errors (zero length sequences).
I blast all my 7600 sequences, but only 20! have a hit against the nr database.
This looks very strange to me, so I pick some sequences, and blast manually from the NCBI website, and obviously this give to me results for every sequence..
so, 2 questions, what could be wrong in my local blast?
and two, I can blast all my sequences using the NCBI webpage, but I can't figure out how to download all the results at once, instead I have to download manually one by one, and I can't do this for 7600 sequences, any workaround?
thanks in advance
Cristian
1 answer
I can see two potential problems.
First, you have formatted the BLAST database using makeblastdb, which is from the newer BLAST+ software package. However, to run BLAST you are using the command blastall, which is from the older version of BLAST. You should be running blastn if you used makeblastdb, or blastall if formatdb was used. Messages about zero-length sequences are not necessarily errors; they may just be warnings.
Second: this is probably not an issue but NCBI uses nr to denote non-redundant protein, nt for non-redundant nucleotide. I just wanted to check that your nr is indeed nucleotide sequences.
The problem is to me obviously the blastn vs nr. Try "blastx" instead.
Log in to answer this question.
What's the blast command you used to blast against your local NR database?
blastall -p blastn -d nr -i ps.fas -o blast_out.txt