added PrimateAI, uploaded source file to s3

score-sets/GRCh37/PrimateAI/README.md

@@ -0,0 +1,7 @@
+Paper at: www.nature.com/articles/s41588-018-0167-z
+
+Using hundreds of thousands of common variants from population sequencing of six non-human primate species, PrimateAI is a trained deep neural network to classify variants' pathogenicity.  It predicts impact purely from sequence and predicts secondary structure information and solvent accessibility information.
+
+From the paper:
+
+"Prediction scores for all ~70 million human missense variants on the hg19/GRCh37 genome build with the human+primate deep learning network (PrimateAI) are publicly hosted (https://basespace.illumina.com/s/cPgCSmecvhb4). For practical application of PrimateAI scores, we recommend a threshold of >0.8 for likely pathogenic classification, <0.6 for likely benign, and 0.6-0.8 as intermediate in genes with dominant modes of inheritance, on the basis of the enrichment of de novo variants in cases as compared to controls (Fig. 3d), and a threshold of >0.7 for likely pathogenic and <0.5 for likely benign in genes with recessive modes of inheritance."

score-sets/GRCh37/PrimateAI/make.sh

@@ -0,0 +1,7 @@
+# use 11th column
+if [ ! -s PrimateAI_scores_v0.2.tsv.gz ]; then
+    wget https://s3.us-east-2.amazonaws.com/pathoscore-data/primateai/PrimateAI_scores_v0.2.tsv.gz
+fi
+zgrep -v "^#" PrimateAI_scores_v0.2.tsv.gz | sed '1d' | sed '1s/^/#/' | sort -k1,1 -k2,2n > PrimateAI_scores_v0.2.tsv
+bgzip -f PrimateAI_scores_v0.2.tsv
+tabix -b 2 -e 2 PrimateAI_scores_v0.2.tsv.gz