Update ShapeIt4 from version 4.1.3 to version 4.2.1 (#92)

* Update ShapeIt4 from version 4.1.3 to version 4.2.1

dockerfiles/README.md

@@ -65,7 +65,7 @@ A farm5-specific docker image, providing dependencies for the iRODS import and b
 - samplevcftozarr - `docker build . --tag=samplevcftozarr:1.0`
 - sampleselectvariants - `docker build . --tag=sampleselectvarians:1.0`
 - whatshap - `docker build . --tag=whatshap:1.0`
-- shapeit4 - `docker build . --tag=shapeit4:4.1.3`
+- shapeit4 - `docker build . --tag=shapeit4:4.2.1`
 - bcftools - `docker build . --tag=bcftools:1.11`
 - import - `docker build . --tag=sangerpathogens/irods:4.1.12`
 

dockerfiles/Shapeit4/Dockerfile

@@ -1,22 +1,31 @@
-FROM bashell/alpine-bash@sha256:965a718a07c700a5204c77e391961edee37477634ce2f9cf652a8e4c2db858ff
+FROM ubuntu:20.04
+# Suppress interactive output from installs
 
-ARG shapeit4_version=4.1.3
+ENV DEBIAN_FRONTEND=noninteractive
+ARG shapeit4_version=4.2.1
 ARG htslib_version=1.11
 ARG boost_version_with_dots=1.74.0
 ARG boost_version=1_74_0
 
-RUN apk update \
- && apk add --no-cache \
- build-base \
- zlib-dev \
- bzip2-dev \
- xz-dev \
- ncurses-dev \
- ca-certificates \
- wget
+RUN apt-get update && \
+    apt-get -y --no-install-recommends install \
+      ca-certificates \
+      libbz2-dev \
+      libcurl4-openssl-dev \
+      liblzma-dev \
+      libssl-dev \
+      libncurses5-dev \
+      autoconf \
+      automake \
+      bzip2 \
+      gcc \
+      g++ \
+      make \
+      wget \
+      zlib1g-dev \
+      libboost-all-dev
 
 WORKDIR /tmp
-
 # Install htslib to get bgzip and tabix
 RUN wget -q https://github.com/samtools/htslib/releases/download/${htslib_version}/htslib-${htslib_version}.tar.bz2 && \
   tar xf htslib-${htslib_version}.tar.bz2 && \
@@ -25,7 +34,7 @@ RUN wget -q https://github.com/samtools/htslib/releases/download/${htslib_versio
   ./configure && \
   make install
 
-RUN wget -q https://dl.bintray.com/boostorg/release/${boost_version_with_dots}/source/boost_${boost_version}.tar.gz && \
+RUN wget -q https://boostorg.jfrog.io/artifactory/main/release/${boost_version_with_dots}/source/boost_${boost_version}.tar.gz && \
   tar xf boost_${boost_version}.tar.gz && \
   rm boost_${boost_version}.tar.gz && \
   cd boost_${boost_version} && \
@@ -36,14 +45,14 @@ RUN wget -q https://github.com/odelaneau/shapeit4/archive/v${shapeit4_version}.t
   tar xf v${shapeit4_version}.tar.gz && \
   rm v${shapeit4_version}.tar.gz && \
   cd shapeit4-${shapeit4_version} && \
-  sed -i s^HTSLIB_INC=\$\(HOME\)/Tools/htslib-1.9^HTSLIB_INC=/usr/local/include/htslib^ makefile && \
-  sed -i s^HTSLIB_LIB=\$\(HOME\)/Tools/htslib-1.9/libhts.a^HTSLIB_LIB=/usr/local/lib/libhts.a^ makefile && \
+  sed -i s^HTSLIB_INC=\$\(HOME\)/Tools/htslib-1.11^HTSLIB_INC=/usr/local/include/htslib^ makefile && \
+  sed -i s^HTSLIB_LIB=\$\(HOME\)/Tools/htslib-1.11/libhts.a^HTSLIB_LIB=/usr/local/lib/libhts.a^ makefile && \
   sed -i s^BOOST_INC=/usr/include^BOOST_INC=/usr/local/include/boost/^ makefile && \
   sed -i s^BOOST_INC=/usr/include^BOOST_INC=/usr/local/include/boost/^ makefile && \
   sed -i s^BOOST_LIB_IO=/usr/lib/x86_64-linux-gnu/libboost_iostreams.a^BOOST_LIB_IO=/usr/local/lib/libboost_iostreams.a^ makefile && \
   sed -i s^BOOST_LIB_PO=/usr/lib/x86_64-linux-gnu/libboost_program_options.a^BOOST_LIB_PO=/usr/local/lib/libboost_program_options.a^ makefile && \
   make && \
-  cp bin/shapeit4 /usr/local/bin
+  cp /tmp/shapeit4-${shapeit4_version}/bin/shapeit4.2 /usr/local/bin/shapeit4
 
 WORKDIR /usr/local/bin
 

tasks/gcp/StatisticalPhasingTasks.wdl

@@ -51,15 +51,15 @@ task ShapeIt4 {
     # TODO - how to handle refence as an option
     ReferenceSequence? reference
 
-    String docker_tag = "us.gcr.io/broad-gotc-prod/malariagen/shapeit4:4.1.3"
+    String docker_tag = "us.gcr.io/broad-gotc-prod/malariagen/shapeit4:4.2.1"
     # Compute Engine always stops preemptible instances after they run for 24 hours
-    Int preemptible_tries = 0
+    Int preemptible_tries = 2
     Int num_cpu = 4
     Float mem_gb = 15
     RunTimeSettings runTimeSettings
   }
 
-  Int disk_size = ceil(size(merged_vcf, "GiB") + size(merged_vcf_index, "GiB") + size(genetic_map, "GiB")) * 20 + 20
+  Int disk_size = ceil(size(merged_vcf, "GiB") + size(merged_vcf_index, "GiB") + size(genetic_map, "GiB")) * 5 + 20
   String output_prefix = sub(region, ":", "_")
   String output_filename = output_prefix + "_" +  project_id + "_phased.vcf.gz"