Blast will only return the information it knows about from the query and the database you're querying against. The output formats are pretty immutable as far as I know.

Not every entry in nr is going to have a gene NAME, nor should it, so why would you expect BLAST to return it? It is sensible to return the accession your query matches against, so that you can look it up and do further analysis on it. It's the annotations attached to the accessions that may give you a clue to the identity of the gene.

I think that with XML output file (-outfmt 5 option I think) in the tag called Hit_def you get the whole FASTA header of the sequence your query had a hit.

I'm not completely sure but you could try. I blasted a sequence against uniprot having a hit with this protein A8MSE8 whose FASTA header is

>tr|A8MSE8|A8MSE8_ARATH Elongation factor 1-alpha OS=Arabidopsis thaliana GN=At5g60390 PE=3 SV=1

and I got this in the XML hit_def tag

tr|A8MSE8|A8MSE8_ARATH Elongation factor 1-alpha OS=Arabidopsis thaliana GN=At5g60390 PE=3 SV=1

So if the gene name is in the FASTA header you could get it

HTH,
Marina

As far as I know blast does not know what part of your query fasta file header is GeneID, clone id or whatever. It simply truncates the first field. So you have choice of reformatting your query Fasta (not recommended unless you do not get unique IDs in your blast output) or try to combine information which you already have in your fasta headers with tabular blast output with a bit of scripting.

Problem is, that fasta headers are far from consistent, may not even contain gene id (i.e. ESTs) so sticking with accessions is way safer.