blast of the contigs

[dina]

Hello,
I've sequenced a genome and received contigs. I'm trying to blast the contigs using megablast installed locally on our computer. For many of the contigs I receive the error message: could not calculate ungapped Karlin0Altshul parameters due to an invalid query sequence or its translation. Please verify the query sequence and/ or filtering options" .
Anybody knows what this means? Thank you !!!

[rglover]

Hi,
How big are your contigs? It would help if you could let us know what parameters/commands you are using to run the blast.
I've sometimes had that error message on 454 reads that consist only of sequence like (AC)n etc.
Cheers,
Rachel

[dina]

blast

Than you for your reply, I'm using the command:
megablast -b -1 -m 9 -i mapping/454AllContigs.fna -d /root/genome_annotation/blast/ncbi_nt/nt>blastout.txt

There are different contigs some are short and others are long

[westerman]

That command line parameter of '-b -1' looks wrong. I will assume that you meant to and are actually using '-b 1' In which case I'll agree with rglover that it is either the composition and/or length of your input sequences that is causing problems. I suspect that we all run into that error message when we work with off-the-sequencer sequences.

It would be helpful if you could post

(a) approximately how many of your sequences are having the problem (i.e., 1%, 10%, 50%?)

(b) the actual nucleotides in one or two of your problem sequences

[dina]

blast

It happens for about 10% of the reads. I guess it is indeed because the quality of the read.
another question about blast: when I run it with this command:megablast, I receive in the end a big html file, part of wfich just says "no hits" for some of the contigs. Is there any way to present it in a more readable format? begining with the best hits (aas in the web) and to show the aligment (the format that I recive when I run blast at the web),
Thanks alot!

[westerman]

Your question "as in the web" is vague. I presume you are talking about NCBI blast but even there you can receive the results back in a variety of different formats. Their default is closer to the '-m 0' format. If you have looked at the other megablast format besides the one you gave (e.g., -m 9) and none of them a good for you then please describe exactly what you want and maybe we can help out.

Personally when looking at the thousands of sequences I usually get with a 2nd gen sequencing project, the '-m 8' (tabular) or '-m 7' options are the ones I use. Formats besides those two become hard to parse and summarize. But your requirements are probably much different than mine.