From a325ef9c17ad9b506204912bde7cbc2a6e1327b5 Mon Sep 17 00:00:00 2001
From: poeli <po-e@lanl.gov>
Date: Sat, 21 May 2022 23:02:27 -0600
Subject: [PATCH] fix gottcha2 SAM format bug

---
 Dockerfile                 | 25 +++++++++++--------------
 ReadbasedAnalysis.wdl      |  2 +-
 ReadbasedAnalysisTasks.wdl |  8 +++++---
 3 files changed, 17 insertions(+), 18 deletions(-)

diff --git a/Dockerfile b/Dockerfile
index b4de3b5..c217704 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -1,15 +1,15 @@
-FROM continuumio/miniconda3:4.8.2
+FROM continuumio/miniconda3:latest
 
 LABEL developer="Po-E Li"
 LABEL email="po-e@lanl.gov"
-LABEL version="1.0.1"
+LABEL version="1.0.3"
 LABEL software="nmdc_taxa_profilers"
 LABEL tags="metagenome, bioinformatics, NMDC, taxonomy"
 
 ENV container docker
 
-RUN apt-get update -y \
-    && apt-get install -y build-essential unzip wget curl gawk \
+RUN apt-get update --allow-releaseinfo-change \
+    && apt-get install -y build-essential \
     && apt-get clean
 
 # add conda channels
@@ -17,20 +17,17 @@ RUN conda config --add channels conda-forge \
     && conda config --add channels bioconda
 
 # install gottcha2
-RUN conda install minimap2 pandas
-RUN wget https://github.com/poeli/GOTTCHA2/archive/2.1.7.tar.gz \
-    && tar -xzf 2.1.7.tar.gz \
-    && cp GOTTCHA2-2.1.7/*.py /usr/local/bin \
-    && rm -rf GOTTCHA2-2.1.7/ 2.1.7.zip
+RUN conda install minimap2 pandas gawk curl
+RUN wget https://github.com/poeli/GOTTCHA2/archive/refs/tags/2.1.8.1.tar.gz \
+    && tar -xzf 2.1.8.1.tar.gz \
+    && cp GOTTCHA2-2.1.8.1/gottcha/scripts/*.py /usr/local/bin \
+    && rm -rf GOTTCHA2-2.1.8.1/ 2.1.8.1.tar.gz
 
 # install kraken2
-RUN conda install kraken2=2.1.0
+RUN conda create -n kraken kraken2=2.1.0
 
 # install centrifuge
-RUN wget https://github.com/DaehwanKimLab/centrifuge/archive/v1.0.4-beta.tar.gz \
-    && tar -xzf v1.0.4-beta.tar.gz \
-    && cd centrifuge-1.0.4-beta \
-    && make install prefix=/usr/local
+RUN conda create -n centrifuge centrifuge=1.0.4_beta
 
 # install krona
 RUN conda install krona \
diff --git a/ReadbasedAnalysis.wdl b/ReadbasedAnalysis.wdl
index 1de443e..ae8ede5 100644
--- a/ReadbasedAnalysis.wdl
+++ b/ReadbasedAnalysis.wdl
@@ -8,7 +8,7 @@ workflow ReadbasedAnalysis {
     String prefix
     String outdir
     Boolean? paired = false
-    String? docker = "microbiomedata/nmdc_taxa_profilers:1.0.2"
+    String? docker = "microbiomedata/nmdc_taxa_profilers:1.0.3"
 
     if (enabled_tools["gottcha2"] == true) {
         call tasks.profilerGottcha2 {
diff --git a/ReadbasedAnalysisTasks.wdl b/ReadbasedAnalysisTasks.wdl
index e745e17..23e0ff1 100644
--- a/ReadbasedAnalysisTasks.wdl
+++ b/ReadbasedAnalysisTasks.wdl
@@ -46,13 +46,14 @@ task profilerCentrifuge {
 
     command <<<
         set -euo pipefail
-
+        conda activate centrifuge
         centrifuge -x ${DB} \
                    -p ${CPU} \
                    -U ${sep=',' READS} \
                    -S ${PREFIX}.classification.tsv \
                    --report-file ${PREFIX}.report.tsv
-        
+        conda deactivate        
+
         ktImportTaxonomy -m 5 -t 2 -o ${PREFIX}.krona.html ${PREFIX}.report.tsv
     >>>
     output {
@@ -84,13 +85,14 @@ task profilerKraken2 {
 
     command <<<
         set -euo pipefail
-        
+        conda activate kraken
         kraken2 ${true="--paired" false='' PAIRED} \
                 --threads ${CPU} \
                 --db ${DB} \
                 --output ${PREFIX}.classification.tsv \
                 --report ${PREFIX}.report.tsv \
                 ${sep=' ' READS}
+        conda deactivate        
 
         ktImportTaxonomy -m 3 -t 5 -o ${PREFIX}.krona.html ${PREFIX}.report.tsv
     >>>