Blast in parallel with edible outfmt

Question

[K.Gee]

Hello Biostars

I know that I can run blast with with "edible" output by using this command:

blastp -query -FASTA.faa - db MYDATABASE -evalue 1e-5 -outfmt "6 sseqid pident evalue bitscore staxids"

If I want to run the same command in parallel, i am using

ls *.fna| parallel -a - blastp -query {} -db MYDATABASE -evalue 1e-5 -outfmt 6 -out {.}.txt

My issue is when I am trying to combine the outfmt in parallel I got an error and I want to know if I can combine those two commands in one?

I tried:

ls *.fna| parallel -a - blastp -query {} -db MYDATABASE -evalue 1e-5 -outfmt "6 sseqid pident evalue bitscore staxids sscinames sskingdoms stitle" -out {.}.txt

This gives me:

Error: Too many positional arguments (1), the offending value: evalue Error: (CArgException::eSynopsis) Too many positional arguments (1), the offending value: evalue

sorry for my terminology, but I'm a new kid on the block

Thank you in advance

[Joe]

I suspect this is just a matter of how bash and parallel are handling the " differently. Have you tried hard-quoting instead with '?

[K.Gee]

I did but doesnt work either :-(, but thanks for your response anyhow!!!

[lieven.sterck]

does the cmdline using only -outfmt 6 work ? (== the problem starts when you add the "6 ...." part ? )

on a side note: I don't see any immediate advantage in using parallel here (if you want to speed it up somehow, run it on multiple threads)

[K.Gee]

Yes with -outfmt 6 works perfectly!!! My issue starts when I am editing the out format

[Joe]

This is almost certainly a quirk of how parallel is treating " I would think, but I can't spot the issue off hand.

[K.Gee]

I tried almost everything before posting my issue. Also I realised that I cannot increase the -num_threads as I am doing in the "normal" blast search, so I guess that it is a bag of the parallel command than a mistake during my typing.

[cpad0112]

see if this is populating correct queries and then try removing dry-run

$ parallel --dry-run  'blastp -query {} -db MYDATABASE -evalue 1e-5  -outfmt "6 sseqid pident evalue bitscore staxids" -out {.}.txt' ::: *.fna

Also check your parallel version.

[K.Gee]

Works too!!! thank you very much!

Answer 1

[ATpoint]

I second the other suggestions of conflicts with how parallel and bash interpret quotations. I generally find it best to write a function for the command and then feed the function into parallel, avoiding exactly what you experience. I do not see a reason to use -a so leaving it out here (untested, based on your code, I do not know the blast syntax):

function blasty {

  blastp \
    -query "${1}" \
    -out "${1%.fna}".txt \
    -db MYDATABASE -evalue 1e-5 -outfmt "6 sseqid pident evalue bitscore staxids sscinames sskingdoms stitle"

}; export -f blasty

ls *.fna | parallel blasty {}

MYDATABASE is worth checking as well, is this a variable or in this case simply a spaceholder for the actual path to the file (or whatever it is, as said not a blast user myself)?

[K.Gee]

Thank you for the response. I got some question regarding your answer. I paste the command at my terminal but it looks like it missing something as I get a blinking ">" Did I have to remove the \ or I have to run the code in different way?. I m sorry for that but I am still learning bash

[ATpoint]

Just changed it, I accidentally wrote blastp rather than blasty in the parallel call, please try again. When I paste the function it works, make sure this database variable is correct.

[K.Gee]

It works . Thanks a lot !!!!! :-)