how to concatenate blast results (m8) via setting threshold of distance between hits

[Lyn Hsiong]

hi, dear guys

I performance blastn (-m 8) using a query file of many sequences, and for each query sequence, the output contains many fragmental hits of significance.

however, these hits have no overlap, and what is interesting is that most gaps < 300bp (much shorter than full-length of the query sequence).

so, how can i concatenate those closely related hits into one via setting a value (e.g 300bp) when these hits match the same subject (different regions), ——also to reduce the number of output hits per query.

for example:

http://www.freeimagehosting.net/bohcp

are there any scripts or tools for this purpose?

all your replies are welcome!

[rhinoceros]

It looks like you're blasting against a protein database and getting hits to CDS (i.e. the gaps represent introns or intergenic regions). If this is the case, you could just extract the ORFs and concatenate them by some simple command like cat. In any case, there must be a one-liner of cat grep cut and awk that would solve your problem..

[Lyn Hsiong]

It looks like you're blasting against a protein database and getting hits to CDS (i.e. the gaps represent introns or intergenic regions). If this is the case, you could just extract the ORFs and concatenate them by some simple command like cat. In any case, there must be a one-liner of cat grep cut and awk that would solve your problem..

thank you for your reply!
actually, what you mentioned may typify one of the cases. then here i performed a tblastx using two genome sequences and found some of the hits per query had close relationship (certain significant colinearity) although each of the component hits seemed to be less related. so If I arbitrarily set a E/score value I think I will lose some informative hits.
here I asked a general question for this kind of application, and I hope there is a versatile scripts...