how to use peptide-lengths parameter?

[kloot]

Hi, thanks for providing this tool! I have trouble figuring out how to pass values with the --peptide-lengths option to mhcflurry-predict-scan. --peptide-lengths 8,9,10 results in error: argument --peptide-lengths: invalid int value: '8,9,10' Various other formats result in other errors - please let me know if this is expected behavior. Thanks in advance!

[timodonnell]

Hi there, try passing each peptide length as a separate argument, eg:

mhcflurry-predict-scan  \
   --sequences MFVFLVLLPLVSSQCVNLTTRTQLPPAYTNSFTRGVYYPDKVFRSSVLHS   \
   --alleles HLA-A*02:01 \
   --peptide-lengths 8 9 10

[kloot]

Thanks for your quick reply, Tim! I can confirm your suggestion works, however

mhcflurry-predict-scan \
--alleles HLA-A*02:01 \
--peptide-lengths 8 9 10 \
--input-format fasta \
input.fasta

throws the same error as in my original message. The parameter combination behaves as expected (works) if I omit --peptide lengths 8 9 10. Odd?

[kloot]

Sorry I was wrong!

[kloot]

Works:

mhcflurry-predict-scan \
--alleles HLA-A*02:01 \
--peptide-lengths 8 9 10 \
--input-format fasta \
input.fasta

Does not work:

mhcflurry-predict-scan \
--alleles HLA-A*02:01 \
--input-format fasta \
--peptide-lengths 8 9 10 \
input.fasta

A little unexpected that parameter order is important - however happy to have made it work! May thanks!

[timodonnell]

Ah i see - yes that is confusing! Probably should let it also accept a separated list as a fix in a future version.

On Fri, Apr 10, 2020 at 10:13 AM kloot notifications@github.com wrote:

Works:

mhcflurry-predict-scan \ --alleles HLA-A*02:01 \ --peptide-lengths 8 9 10 \ --input-format fasta \ input.fasta

Does not work:

mhcflurry-predict-scan \ --alleles HLA-A*02:01 \ --input-format fasta \ --peptide-lengths 8 9 10 \ input.fasta

A little unexpected that parameter order is important - however happy to have made it work! May thanks!

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/openvax/mhcflurry/issues/160#issuecomment-612046197, or unsubscribe https://github.com/notifications/unsubscribe-auth/AADSHODJPXBXIBS2EZM5YUTRL4SPVANCNFSM4MFOMULA .

[kloot]

Good idea, that was my first assumption as it's consistent with --alleles. Thank you again for making this software available. I don't know how to add a 'question' flag, as it's not truly an 'issue' ;-)

[JamieHeather]

Maybe allowing a range might also be helpful? E.g. 8-10, or 8:10 or something.

[kloot]

If I'm allowed an off-topic suggestion for future versions: Permit lower-case characters in the input sequence and retain upper/lower case in the output. Lower-case letters are FASTA-conform and will make it much easier to track mutant residues in neopeptides in the mhcflurry output.

[timodonnell]

Both sound like good ideas. Leaving this ticket open as a reminder to do these for the next version (or feel free to open separate issues)

[kloot]

Thank you in advance Tim!