Any interest for tblastx for comparison?


Dear all,
We use tblastx (translating query to all reading frames and compare it with a nucleotide database translated to all reading frames = computing time is equivalent to 36 x blastp) to find remote relatives for virus isolates that we have found in environmental samples. A while ago I asked Daniel if tblastx could be included in Megan. Daniel already has blastx functionality. He replied that tblastx would require quite a bit of work and he never had this request before. I think that without tblastx Megan is already a very useful piece of software, and I do not want to bother Daniel with this, unless this functionality would be useful for other people.

So, please let me/Daniel know if you would like to have this functionality. If not I, am ok with it, but if you do want to use it, and if Daniel can implement, I would definitely use it.


If you save the tblastx alignments in tab format then MEGAN should be able to deal with the file (choose blast tab as format and choose blastp as the mode).
You won’t be able to inspect the alignments, but most of all other features should still work.

Also, could you please send me a small example (alignments and reads) and I will look into adding tblastx support. (I won’t be able to work on it immediately, but it is something that I can look into on a long train ride or airport layover).


OK I will have a look tblastx tabular as soon as we have finished setting up a faster system (should be next week).
Now I can cheat already by changing the first line in the tblastx text output from TBLASTX to BLASTX and tell Megan that I have run a blastx, but it lacks some functionality, e.g. I cannot use a mapping file.


That is a good idea. By changing the first line to BLASTX, MEGAN will be 90% fooled and most stuff should work.
My guess is the only feature that will break is the alignment viewer, because it will assume that the alignment coordinates of the subject is in amino-acids, but they will be in nucleotides.
In want sense do you mean “cannot use a mapping file”? As this is a DNA to DNA alignment, you need use the file nucl_acc2tax-May2017.abin, but that should work.


Dear Daniel,
Using the tblastx to blastx change of line 1 also worked with the nucl_acc2tax-May2017.abin mapping file, using a script. An earlier attempt with mapping file, using the graphical interface, did not work, but I have not tested again.

If someone wants to include the following 2 lines in a bash script (I used a Refseq database for blast, identifiers start with “ref|”, $ID is the first part of the filename, $MAPfile is the full path to nucl_acc2tax-May2017.abin):
perl -pi -e ‘s/TBLASTX/BLASTX/’ $ID.tx.out
/opt/megan/tools/blast2rma -i $ID.tx.out -f BlastText -bm BlastX -r $ID.fna -o $ID.txmap.rma6 -c -lg -a2t $MAPfile -atags “gb| ref| refseq|”

For me this does the job, and if I want to check an alignment I can do that in the blast output, so I do not have another request than ‘just’ to include the option that it can be used for tblastx, too.

Daniel, if you still want to have a fasta-tblastx_out combination to play with, I can send it to you directly


Yes, please send me an example and I will look into it.