Reciprocal blast help

[Shorash]

I have 2 datasets (Dataset A and dataset B) in which I have reciprocally allocated as BLASTx hits against one another. However I am having difficulties in identifying those contigs that are hits to one another from both datasets. I have been able to export the BLAST results into sequence tables but I'm not sure how I can identify the reciprocal top hits of one another for a large number of contigs (160,000). Here is an example of how it looks in an Excel spreadsheet:

aaaaaDataset Aaaaaaaaaaaaaaaaa Dataset B
Contig123=Contig789_1 Contig789=111Contig123_1
Contig456=72Contig221 Contig221=Contig456_3
Contig777=43Contig954 Contig954=3Contig1561_1


In the example above you can see that the results or hit from each file have characters on the beginning and sometimes on the end of each corresponding hit making it hard to compare using excel formulas. In the example, the first two rows are the ones I'm interested in extracting as they have hit the same contig in both datasets, unlike row 3 which do not match.

Any help would be greatly appreciated!

[bio_boris]

What have you come up with so far? Do you have any scripts to use or code that you have tried to create?

Some resources

http://ged.msu.edu/angus/tutorials/

http://ged.msu.edu/angus/tutorials/reciprocal-blast.html

http://seqanswers.com/forums/showthread.php?t=20652

[someperson]

Not really a direct answer to your question, but a tip for a tool that probably already does what you want:
I'm mostly using the tool proteinortho (curren version proteinortho5) for reciprocal blast analyses.
A strong advantage of this tool is, that it does not only direct orthologs via direct reziprocal blast, but can also list the respective paralogs and group them into "orthologeous groups".
links:

https://www.bioinf.uni-leipzig.de/Software/proteinortho/

http://www.biomedcentral.com/1471-2105/12/124

[Shorash]

What have you come up with so far? Do you have any scripts to use or code that you have tried to create?

Some resources

http://ged.msu.edu/angus/tutorials/

http://ged.msu.edu/angus/tutorials/reciprocal-blast.html

http://seqanswers.com/forums/showthread.php?t=20652

I haven't managed to create any scripts or codes. I've been manually looking at specific genes of interest but it would be great to be able to do all of them at once.

[Shorash]

Not really a direct answer to your question, but a tip for a tool that probably already does what you want:
I'm mostly using the tool proteinortho (curren version proteinortho5) for reciprocal blast analyses.
A strong advantage of this tool is, that it does not only direct orthologs via direct reziprocal blast, but can also list the respective paralogs and group them into "orthologeous groups".
links:

https://www.bioinf.uni-leipzig.de/Software/proteinortho/

http://www.biomedcentral.com/1471-2105/12/124

Great thanks for that, I'll give this a try.

[rhinoceros]

how I can identify the reciprocal top hits of one another for a large number of contigs (160,000).

I would go for tabular blast output and first sort for best hits. So then, depending how you did your blasts, you can have e.g. two best-hit sorted output files with query in the first column and subject in the second. One option would be to cut columns 1-2 and switch the the order in one file and then cat it with the other file. Then you'd sort based on column 1 and only output the lines where uniq -c is 2. I'm sure there's an awk one-liner for this too..