First pass on models_class1_allele_specific_single_kim2014_only

downloads-generation/models_class1_allele_specific_single_kim2014_only/GENERATE.sh

+#!/bin/bash
+
+if [[ $# -eq 0 ]] ; then
+    echo 'WARNING: This script is intended to be called with additional arguments to pass to mhcflurry-class1-allele-specific-cv-and-train'
+    echo 'At minimum you probably want to pass --dask-scheduler <IP:PORT> as training many models on one node is extremely '
+    echo 'slow.'
+fi
+
+set -e
+set -x
+
+DOWNLOAD_NAME=models_class1_allele_specific_single_kim2014_only
+SCRATCH_DIR=/tmp/mhcflurry-downloads-generation
+SCRIPT_ABSOLUTE_PATH="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)/$(basename "${BASH_SOURCE[0]}")"
+SCRIPT_DIR=$(dirname "$SCRIPT_ABSOLUTE_PATH")
+
+mkdir -p "$SCRATCH_DIR"
+rm -rf "$SCRATCH_DIR/$DOWNLOAD_NAME"
+mkdir "$SCRATCH_DIR/$DOWNLOAD_NAME"
+
+# Send stdout and stderr to a logfile included with the archive.
+exec >  >(tee -ia "$SCRATCH_DIR/$DOWNLOAD_NAME/LOG.txt")
+exec 2> >(tee -ia "$SCRATCH_DIR/$DOWNLOAD_NAME/LOG.txt" >&2)
+
+# Log some environment info
+date
+pip freeze
+git rev-parse HEAD
+git status
+
+cd $SCRATCH_DIR/$DOWNLOAD_NAME
+
+mkdir models
+
+cp $SCRIPT_DIR/models.py $SCRIPT_DIR/imputer.json .
+python models.py > models.json
+
+time mhcflurry-class1-allele-specific-cv-and-train \
+    --model-architectures models.json \
+    --imputer-description imputer.json \
+    --train-data "$(mhcflurry-downloads path data_kim2014)/bdata.2009.mhci.public.1.txt" \
+    --test-data "$(mhcflurry-downloads path data_kim2014)/bdata.2013.mhci.public.blind.1.txt" \
+    --min-samples-per-allele 70 \
+    --out-cv-results cv.csv \
+    --out-production-results production.csv \
+    --out-models models \
+    --verbose \
+    "$@"
+
+cp $SCRIPT_ABSOLUTE_PATH .
+tar -cjf "../${DOWNLOAD_NAME}.tar.bz2" *
+
+echo "Created archive: $SCRATCH_DIR/$DOWNLOAD_NAME.tar.bz2"

downloads-generation/models_class1_allele_specific_single_kim2014_only/README.md

+# Class I allele specific models (single) trained and tested in Kim 2014 dataset
+
+This is a reimplementation of the analysis in [Predicting Peptide-MHC Binding Affinities With Imputed Training Data](http://biorxiv.org/content/early/2016/05/22/054775).
+

downloads-generation/models_class1_allele_specific_single_kim2014_only/imputer.json

+{
+    "imputation_method_name": "mice",
+    "n_burn_in": 5,
+    "n_imputations": 50,
+    "n_nearest_columns": 25,
+    "min_observations_per_peptide": 5,
+    "min_observations_per_allele": 100 
+}

downloads-generation/models_class1_allele_specific_single_kim2014_only/models.py

+import sys
+from mhcflurry.class1_allele_specific.train import HYPERPARAMETER_DEFAULTS
+import json
+
+models = HYPERPARAMETER_DEFAULTS.models_grid(
+    #impute=[False, True],
+    impute=[False],
+    activation=["tanh"],
+    layer_sizes=[[12], [64], [128]],
+    embedding_output_dim=[8, 32, 64],
+    dropout_probability=[0, .1, .25],
+    # fraction_negative=[0, .1, .2],
+    n_training_epochs=[250])
+
+sys.stderr.write("Models: %d\n" % len(models))
+print(json.dumps(models, indent=4))