Skip to content
Snippets Groups Projects
Select Git revision
  • publish
  • master default protected
  • arch2
  • recover_lost_data
  • manuscript
  • improved-allele-parsing
  • 202010-phospho
  • gh-pages
  • pre-2.0
  • 1.7.0c
  • 3-flank-variant
  • 1.7.0b
  • 1.7.0
  • pre-presentation-refactor2
  • 20200121_pre_presentation_refactor
  • issue-149
  • 201911
  • ligandome-prediction
  • docs-pan-motifs
  • v1.3
  • pre-2.1
  • v2.0.1
  • v2.0.0
  • pre-2.0
  • 1.6.1
  • pre-1.7.0
  • pre_presentation_refactor
  • 1.6.0
  • 1.4.0
  • pre-1.4.0
  • pan-dev1
  • 1.2.2
  • 1.2.0
  • pre-1.2.1
  • pre-1.2
  • pre-1.1
  • 1.0.0
  • pre-1.0
  • 0.9.2
  • 0.9.1
40 results
Blame Permalink
commandline_tutorial.rst 1.48 KiB

Command-line usage

Downloading models

Most users will use pre-trained MHCflurry models that we release. These models are distributed separately from the source code and may be downloaded with the following command:

We also release other "downloads," such as curated training data and some experimental models. To see what you have downloaded, run:

mhcflurry-predict

The mhcflurry-predict command generates predictions from the command-line. It defaults to using the pre-trained models you downloaded above but this can be customized with the --models argument. See mhcflurry-predict -h for details.

The predictions returned are affinities (KD) in nM. The prediction_low and prediction_high fields give the 5-95 percentile predictions across the models in the ensemble. The predictions above were generated with MHCflurry |version|.

Your exact predictions may vary slightly from these (up to about 1 nM) depending on the Keras backend in use and other numerical details. Different versions of MHCflurry can of course give results considerably different from these.

You can also specify the input and output as CSV files. Run mhcflurry-predict -h for details.

Fitting your own models