Merge pull request #1 from vadimzalunin/master

pipeline for 3k contigs taxonomy using mash

src/taxonomy/Dockerfile

@@ -0,0 +1,16 @@
+FROM debian:stretch
+
+RUN apt-get update
+RUN apt-get install -qy wget
+
+RUN wget -q https://repo.anaconda.com/archive/Anaconda3-2019.10-Linux-x86_64.sh && bash Anaconda3-2019.10-Linux-x86_64.sh -b -p /root/anaconda 
+ENV PATH="/root/anaconda/bin:${PATH}"
+RUN eval $(conda shell.bash hook) && conda init
+
+RUN conda install -y -c bioconda -c conda-forge snakemake
+
+RUN conda install google-cloud-storage
+
+COPY bloom_filter bloom_check /usr/bin/
+
+RUN wget -q https://github.com/marbl/Mash/releases/download/v2.2/mash-Linux64-v2.2.tar && tar xf mash-Linux64-v2.2.tar && mv mash-Linux64-v2.2/mash /usr/bin/

src/taxonomy/Snakefile

@@ -0,0 +1,23 @@
+from snakemake.remote.GS import RemoteProvider as GSRemoteProvider
+GS = GSRemoteProvider()
+
+configfile: 'config.yaml'
+
+SAMPLES, *_ = GS.glob_wildcards(config['bucket'] + '/'+ config['dir']+'/{acc}.realign.local.fa')
+msh=GS.remote(config["ref_sketch"])
+
+rule sketch:
+    input: fasta="3k/{acc}.realign.local.fa"
+    output: msh="3k/{acc}.realign.local.fa.msh"
+    run: shell("mash sketch -i {input.fasta}")
+
+rule dist:
+    input: sketch="3k/{acc}.realign.local.fa.msh", msh=msh
+    output: dist="3k/{acc}.realign.local.fa.dist"
+    threads: 6
+    shell: '''
+mash dist -v 0.05 -p 6 {input.msh} {input.sketch} | sed -rn 's/\\t([^0][0-9]+)\\/([0-9]+)$/\\t\\1\\t\\2/p' | sort -nrk5 > {output.dist}
+'''
+
+rule check_all:
+    input: expand("3k/{acc}.realign.local.fa.dist", acc=SAMPLES) 

src/taxonomy/config.yaml

@@ -0,0 +1,3 @@
+bucket: viral2_test_bucket
+dir: 3k
+ref_sketch: viral2_test_bucket/ref_viruses_rep_genomes_v5.fasta.msh

src/taxonomy/mash_tax.sh

@@ -0,0 +1,4 @@
+#!/bin/bash -eu
+
+bucket=$1
+snakemake --rerun-incomplete -p -j 750 --kubernetes --container-image us.gcr.io/strides-sra-hackathon-data/test_pipeline:v0.4 --default-remote-provider GS --default-remote-prefix $bucket --latency-wait 60 --keep-going --restart-times 3 --nolock check_all