Ug master

.github/workflows/build-docker.yml

@@ -0,0 +1,24 @@
+name: Build HLA-LA Docker image
+
+on:
+  workflow_dispatch:
+    inputs:
+      image-tag:
+        description: "Docker image tag, If omitted will use github short sha"
+        required: false
+        type: string
+
+permissions:
+  id-token: write # Required for assuming an AWS role
+  contents: read # Required for actions/checkout
+
+jobs:
+  build:
+    uses: Ultimagen/ugbio-utils/.github/workflows/docker-build-push.yml@main
+    secrets: inherit
+    with:
+      git-repo: ${{ github.repository }}
+      git-ref: ${{ github.ref_name }}
+      docker-image: hla_la
+      docker-tag: ${{ inputs.image-tag }}
+      base-branch: ug-master

.github/workflows/trivy-scan.yml

@@ -0,0 +1,19 @@
+name: Trivy scan
+on:
+  push:
+    branches:
+      - ug-master
+  pull_request:
+jobs:
+  trivy-scan-fs:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: aquasecurity/trivy-action@0.32.0
+        with:
+          scan-type: 'fs'
+          scan-ref: './'
+          format: 'table'
+          exit-code: '1'
+          severity: 'CRITICAL,HIGH'

Dockerfile

@@ -0,0 +1,89 @@
+# work from latest LTS ubuntu release
+FROM ubuntu:22.04
+
+# set the environment variables
+ENV hla_la_version=1.0.4 \
+    samtools_version=1.21 \
+    bwa_version=0.7.19 \
+    picard_version=2.25.7 \
+    bamtools_version=2.5.2
+
+# install required system packages
+RUN apt-get update -y && apt-get install --no-install-recommends -y \
+    build-essential \
+    curl \
+    unzip \
+    zlib1g-dev \
+    libncurses5-dev \
+    libncursesw5-dev \
+    libnss-sss \
+    libbz2-dev \
+    liblzma-dev \
+    vim \
+    less \
+    libcurl4-openssl-dev \
+    wget \
+    libz-dev \
+    openjdk-11-jre \
+    libboost-all-dev \
+    cmake \
+    libjsoncpp-dev \
+ && apt-get clean && rm -rf /var/lib/apt/lists/*
+
+# install samtools
+WORKDIR /usr/local/bin
+RUN wget https://github.com/samtools/samtools/releases/download/${samtools_version}/samtools-${samtools_version}.tar.bz2 \
+ && tar -xjf samtools-${samtools_version}.tar.bz2 \
+ && cd samtools-${samtools_version} \
+ && ./configure \
+ && make \
+ && make install
+
+# install picard
+RUN mkdir -p /usr/local/bin/picard && \
+    wget -O /usr/local/bin/picard/picard.jar https://github.com/broadinstitute/picard/releases/download/${picard_version}/picard.jar && \
+    chmod 0644 /usr/local/bin/picard/picard.jar && \
+    echo '#!/bin/bash\nexec java -jar /usr/local/bin/picard/picard.jar "$@"' > /usr/local/bin/run-picard && \
+    chmod +x /usr/local/bin/run-picard
+
+# install bwa
+RUN curl -SL https://github.com/lh3/bwa/archive/v${bwa_version}.zip -o v${bwa_version}.zip \
+ && unzip v${bwa_version}.zip \
+ && cd bwa-${bwa_version} \
+ && make \
+ && ln -s /usr/local/bin/bwa-${bwa_version}/bwa /usr/local/bin/bwa
+
+# install bamtools
+RUN wget https://github.com/pezmaster31/bamtools/archive/v${bamtools_version}.zip \
+ && unzip v${bamtools_version}.zip \
+ && mkdir -p bamtools-${bamtools_version}/build \
+ && cd bamtools-${bamtools_version}/build \
+ && cmake -DCMAKE_INSTALL_PREFIX=/usr/local/bin/bamtools-${bamtools_version} .. \
+ && make \
+ && make install
+
+
+RUN mkdir -p /usr/local/bin/HLA-LA/bin \
+    /usr/local/bin/HLA-LA/src \
+    /usr/local/bin/HLA-LA/obj \
+    /usr/local/bin/HLA-LA/temp \
+    /usr/local/bin/HLA-LA/working \
+    /usr/local/bin/HLA-LA/graphs
+COPY . /usr/local/bin/HLA-LA/src/
+RUN sed -i 's@\$(BAMTOOLS_PATH)/lib64@\$(BAMTOOLS_PATH)/lib@' /usr/local/bin/HLA-LA/src/makefile \
+ && make -C /usr/local/bin/HLA-LA/src all BOOST_PATH=/usr/include/boost BAMTOOLS_PATH=/usr/local/bin/bamtools-${bamtools_version}
+
+# modify paths.ini for hla-la
+RUN sed -i 's@samtools_bin=@samtools_bin=/usr/local/bin/samtools@' /usr/local/bin/HLA-LA/src/paths.ini \
+ && sed -i 's@bwa_bin=@bwa_bin=/usr/local/bin/bwa@' /usr/local/bin/HLA-LA/src/paths.ini \
+ && sed -i 's@picard_sam2fastq_bin=.*@picard_sam2fastq_bin=/usr/local/bin/picard/picard.jar@' /usr/local/bin/HLA-LA/src/paths.ini
+
+# set PATH
+ENV PATH="/usr/local/bin/HLA-LA/bin:/usr/local/bin/HLA-LA/src:$PATH"
+
+RUN groupadd app && useradd -m app -g app
+RUN chown -R app:app /opt /usr/local/bin
+
+USER app
+
+WORKDIR /home/app

HLA-LA.cpp

@@ -756,7 +756,7 @@ int main(int argc, char *argv[]) {
 			assert(arguments.count("longReads"));
 
 			longReads = arguments.at("longReads");
-			assert((longReads == "0") || (longReads == "ont2d") || (longReads == "pacbio"));
+			assert((longReads == "0") || (longReads == "ont2d") || (longReads == "pacbio") || (longReads == "ultimagen"));
 			if(longReads == "0")
 			{
 				longReads = "";

HLA-LA.pl

@@ -91,9 +91,9 @@
 	die unless(-e '../exonkMerExtraction/exonCoordinates_manual.txt.forExtraction');	
 }
 
-unless((not $longReads) or ($longReads eq 'ont2d') or ($longReads eq 'pacbio'))
+unless((not $longReads) or ($longReads eq 'ont2d') or ($longReads eq 'pacbio') or ($longReads eq 'ultimagen'))
 {
-	die "Please specify --longReads ont2d or --longReads pacbio";
+	die "Please specify --longReads ont2d or --longReads pacbio or --longReads ultimagen";
 }
 
 my %paths_ini;

HLAtypeinference_validation.pl

@@ -450,7 +450,11 @@
 		if($reference_lookup_ID =~ /_Nanopore/i)
 		{
 			$reference_lookup_ID =~ s/_Nanopore//i;			
-		}		
+		}
+		if($reference_lookup_ID =~ /_Ultimagen/i)
+		{
+			$reference_lookup_ID =~ s/_Ultimagen//i;			
+		}
 		unless(exists $reference_data{$reference_lookup_ID})
 		{
 			$missing_reference_data{$reference_lookup_ID}{$locus}++;

README.md

@@ -154,7 +154,7 @@ f you work with CRAM files and receive error messages like `Unable to fetch refe
 
 ### Long reads
 
-If you want to carry out genotyping from a long-reads BAM, please specify the parameter `--longReads ont2d` (for Oxford Nanopore reads) or `--longReads pacbio` (for PacBio reads). Of note, typing will still be carried out at G group resolution. This feature is experimental; accuracy may not match short-read-based typing.
+If you want to carry out genotyping from a long-reads BAM, please specify the parameter `--longReads ont2d` (for Oxford Nanopore reads) or `--longReads pacbio` (for PacBio reads) or `--longReads ultimagen` (for Ultima Genomics reads which are not long reads, but unpaired reads). Of note, typing will still be carried out at G group resolution. This feature is experimental; accuracy may not match short-read-based typing.
 
 ### Assembly typing
 

hla/HLATyper.cpp

@@ -1029,7 +1029,7 @@ void HLATyper::HLATypeInference(const std::vector<mapper::reads::oneReadPair>& r
 
 	// read alignment statistics
 
-	int minAlignmentLength_unpaired = 1000;
+	int minAlignmentLength_unpaired = (longReadsMode == "ultimagen") ? 50 : 1000;
 	int alignmentStats_strandsValid = 0;
 	int alignments_paired_perfect = 0;
 	int alignments_paired_oneReadPerfect = 0;

mapper/bwa/BWAmapper.cpp

@@ -123,7 +123,7 @@ void BWAmapper::mapUnpaired(std::string indexedReferenceGenome, std::string FAST
 
 void BWAmapper::mapLong(std::string indexedReferenceGenome, std::string FASTQ, std::string outputBAM, bool withA, std::string longMode)
 {
-	assert((longMode == "pacbio") || (longMode == "ont2d"));
+	assert((longMode == "pacbio") || (longMode == "ont2d") || (longMode == "ultimagen"));
 
 	assert(Utilities::fileExists(bwa_bin));
 	assert(Utilities::fileExists(samtools_bin));
@@ -150,7 +150,7 @@ void BWAmapper::mapLong(std::string indexedReferenceGenome, std::string FASTQ, s
 	assert(threads_minus_1 >= 0);
 
 	//std::string bwa_mapping_cmd = bwa_bin + " mem " + indexedReferenceGenome + " " + FASTQ1 + " " + FASTQ2 + " > " + outputSAM;
-	std::string bwa_mapping_cmd = bwa_bin + " mem -t" + Utilities::ItoStr(threads) + " -x " + longMode + " -M " + with_a + indexedReferenceGenome + " " + FASTQ + " | " + samtools_bin + " view -@ " + Utilities::ItoStr(threads_minus_1) + " -Sb - > " + outputUnsorted;
+	std::string bwa_mapping_cmd = bwa_bin + " mem -t" + Utilities::ItoStr(threads) + " -x " + ((longMode =="ultimagen")? "ont2d" : longMode)+ " -M " + with_a + indexedReferenceGenome + " " + FASTQ + " | " + samtools_bin + " view -@ " + Utilities::ItoStr(threads_minus_1) + " -Sb - > " + outputUnsorted;
 	std::cerr << bwa_mapping_cmd << "\n" << std::flush;
 	int retCode = std::system(bwa_mapping_cmd.c_str());
 	if(retCode != 0)