Dev0420 (#6)

* virsorter, bacphlip and mmseqs taxa fix

.github/workflows/docker.yml

@@ -3,7 +3,7 @@ name: Docker
 on:
   push:
     tags:
-      # - 'v*'
+      - 'v*'
       - 'docker*'
 
 jobs:

conf/modules.config

@@ -57,11 +57,11 @@ process {
     }
 
     withLabel: viroprofiler_replicyc {
-        container = 'denglab/viroprofiler-replicyc:v0.2'
+        container = 'denglab/viroprofiler-replicyc:v0.1'
     }
 
     withLabel: viroprofiler_taxa {
-        container = 'denglab/viroprofiler-taxa:v0.2'
+        container = 'denglab/viroprofiler-taxa:base'
     }
 
     withLabel: viroprofiler_virsorter2 {

custom.config

@@ -54,7 +54,7 @@ process {
         cpus = { check_max( 1 * task.attempt, 'cpus') }
         memory = { check_max( 4.GB * task.attempt, 'memory') }
     }
-    
+
     withName: CHECKV {
         cpus = { check_max( 1 * task.attempt, 'cpus') }
         memory = { check_max( 20.GB * task.attempt, 'memory') }
@@ -245,4 +245,4 @@ def check_max(obj, type) {
             return obj
         }
     }
-}
+}

docker/viroprofiler-base/env_base.yml

@@ -15,5 +15,5 @@ dependencies:
     - wget
     - pigz
     - coverm
-    - mmseqs2=13
+    - mmseqs2=13.45111
     - unzip

docker/viroprofiler-taxa/Dockerfile

@@ -10,21 +10,32 @@ RUN --mount=type=cache,target=/opt/conda/pkgs micromamba install -n base -f /tmp
 
 # Activate conda env during docker build
 ARG MAMBA_DOCKERFILE_ACTIVATE=1
+
+
+RUN cd /opt/conda/lib/python3.8/site-packages/vcontact2/data && \
+    /opt/conda/bin/wget https://bitbucket.org/MAVERICLab/vcontact2/raw/6089ed83fd03a844de3140c442863c7b29881d43/vcontact2/data/ViralRefSeq-prokaryotes-v211.Merged-reference.csv && \
+    /opt/conda/bin/wget https://bitbucket.org/MAVERICLab/vcontact2/raw/6089ed83fd03a844de3140c442863c7b29881d43/vcontact2/data/ViralRefSeq-prokaryotes-v211.faa.gz && \
+    /opt/conda/bin/wget https://bitbucket.org/MAVERICLab/vcontact2/raw/6089ed83fd03a844de3140c442863c7b29881d43/vcontact2/data/ViralRefSeq-prokaryotes-v211.protein2contig.csv
+
 # Set PATH manually, because nextflow doesn't activate base env by default.
-ENV PATH=/opt/MMseqs2/build/bin:/opt/conda/bin:$PATH
+# ENV PATH=/opt/MMseqs2/build/bin:/opt/conda/bin:$PATH
+ENV PATH=/opt/mmseqs/bin:/opt/MMseqs2/build/bin:/opt/conda/bin:$PATH
 
 RUN cd /opt && \
-    git clone --depth=1 https://github.com/soedinglab/MMseqs2.git && \
+    git clone https://github.com/soedinglab/MMseqs2.git && \
     cd /opt/MMseqs2 && \
+    git checkout 3b9cf88 && \
     mkdir build && \
     cd build && \
     cmake -DHAVE_SANITIZER=1 -DCMAKE_BUILD_TYPE=ASan -DCMAKE_INSTALL_PREFIX=. .. && \
-    make -j 4 && \
+    make -j 1 && \
     make install
 
 RUN micromamba clean --all --yes
 
-RUN cd /opt/conda/lib/python3.8/site-packages/vcontact2/data && \
-    /opt/conda/bin/wget https://bitbucket.org/MAVERICLab/vcontact2/raw/6089ed83fd03a844de3140c442863c7b29881d43/vcontact2/data/ViralRefSeq-prokaryotes-v211.Merged-reference.csv && \
-    /opt/conda/bin/wget https://bitbucket.org/MAVERICLab/vcontact2/raw/6089ed83fd03a844de3140c442863c7b29881d43/vcontact2/data/ViralRefSeq-prokaryotes-v211.faa.gz && \
-    /opt/conda/bin/wget https://bitbucket.org/MAVERICLab/vcontact2/raw/6089ed83fd03a844de3140c442863c7b29881d43/vcontact2/data/ViralRefSeq-prokaryotes-v211.protein2contig.csv
+
+# COPY ./docker/viroprofiler-taxa/mmseqs /opt/conda/bin
+# RUN chmod +x /opt/conda/bin/mmseqs
+
+RUN cd /opt && \
+    wget https://mmseqs.com/latest/mmseqs-linux-avx2.tar.gz; tar xvfz mmseqs-linux-avx2.tar.gz; rm mmseqs-linux-avx2.tar.gz

docker/viroprofiler-taxa/env_taxa.yml

@@ -9,7 +9,7 @@ dependencies:
     - conda-forge::cmake
     - conda-forge::gcc
     - conda-forge::cxx-compiler
-    - bioconda::mmseqs2
+    # - bioconda::mmseqs2=13.45111
     - bioconda::vcontact2=0.11.1
     - bioconda::diamond=2.0.6
     - bioconda::mcl=14.137

docker/viroprofiler-virsorter2/Dockerfile

@@ -15,5 +15,7 @@ COPY ./docker/viroprofiler-virsorter2/env_*.yml /tmp
 # Install packages
 RUN --mount=type=cache,target=/opt/conda/pkgs micromamba install -n base -f /tmp/env_virsorter2.yml -y
 
+RUN apt-get update && apt-get install git -y && cd / && git clone https://github.com/jiarong/VirSorter2.git && cd VirSorter2 && pip install -e .
+
 # clean
-RUN micromamba clean --all --yes
+#RUN micromamba clean --all --yes

docker/viroprofiler-virsorter2/env_virsorter2.yml

@@ -3,8 +3,20 @@ channels:
     - conda-forge
     - bioconda
 dependencies:
+    - python>=3.6
+    - scikit-learn=0.22.1
+    - imbalanced-learn
+    - pandas
+    - seaborn
+    - hmmer==3.3
+    - prodigal
+    - screed
+    - ruamel.yaml
+    - snakemake>=5.18,<=5.26
+    - click
+    - mamba
     - conda-forge::procps-ng
-    - virsorter=2.2.4
     - seqkit
     - csvtk
-    - conda-forge::glpk
+    - conda-forge::glpk
+

modules/local/replicyc.nf

@@ -34,7 +34,7 @@ process BACPHLIP {
 
     output:
     // TODO nf-core: Named file extensions MUST be emitted for ALL output channels
-    path "*.bacphlip", emit: repcyc_ch
+    path "*.bacphlip", emit: bacphlip_ch
     path "*.hmmsearch.tsv"
     path "versions.yml", emit: versions
     // TODO nf-core: List additional required output channels/values here
@@ -73,7 +73,7 @@ process REPLIDEC {
     path(contigs)
 
     output:
-    path("out_replidec.tsv"), emit: gene_fna_ch
+    path("out_replidec.tsv"), emit: replidec_ch
     path "versions.yml", emit: versions
 
     when:

modules/local/taxonomy.nf

@@ -30,7 +30,7 @@ process TAXONOMY_VCONTACT {
 }
 
 process TAXONOMY_MMSEQS {
-    label "viroprofiler_taxa"
+    label "viroprofiler_base"
 
     input:
     path contigs
@@ -46,7 +46,7 @@ process TAXONOMY_MMSEQS {
     """
     # Run mmseqs taxonomy
     mmseqs createdb $contigs qry
-    mmseqs taxonomy qry ${params.db}/taxonomy/mmseqs_vrefseq/refseq_viral mmseqsTaxaRst tmp --tax-lineage 1 --majority 0.4 --vote-mode 1 --lca-mode 3 --orf-filter 1 --threads $task.cpus
+    mmseqs taxonomy qry ${params.db}/taxonomy/mmseqs_vrefseq/refseq_viral mmseqsTaxaRst tmp --tax-lineage 1 --majority 0.4 --vote-mode 1 --lca-mode 3 --orf-filter 0 --threads $task.cpus
 
     # report
     mmseqs createtsv qry mmseqsTaxaRst mmseqsTaxaRst.tsv
@@ -74,7 +74,7 @@ process TAXONOMY_MERGE {
 
     when:
     task.ext.when == null || task.ext.when
-    
+
     """
     parse_vContact2_vc.py -i $taxa_vc -o taxa_vc2 -a $params.assembler
     parse_mmseqsTaxa.py -i $taxa_mmseqs -o taxa_mmseqs -u "" -s $params.taxa_db_source

workflows/viroprofiler.nf

@@ -272,13 +272,15 @@ workflow VIROPROFILER {
         if ( params.replicyc == "bacphlip" ) {
             BACPHLIP (vContigs_and_vMAGs)
             ch_versions = ch_versions.mix(BACPHLIP.out.versions)
+            ch_replicyc = BACPHLIP.out.bacphlip_ch
         } else if ( params.replicyc == "replidec" ) {
             REPLIDEC (vContigs_and_vMAGs)
             ch_versions = ch_versions.mix(REPLIDEC.out.versions)
+            ch_replicyc = REPLIDEC.out.replidec_ch
         }
 
         // TreeSummarizedExperiment
-        RESULTS_TSE (ABUNDANCE.out.ab_count_ch, ABUNDANCE.out.ab_tpm_ch, ABUNDANCE.out.ab_covfrac_ch, TAXONOMY_MERGE.out.taxa_mmseqs_ch, CHECKV.out.checkv2vContigs_ch, VIRSORTER2.out.vs2_score_ch, VIBRANT.out.vibrant_quality_ch, DVF.out.dvf2vContigs_ch, BACPHLIP.out.repcyc_ch)
+        RESULTS_TSE (ABUNDANCE.out.ab_count_ch, ABUNDANCE.out.ab_tpm_ch, ABUNDANCE.out.ab_covfrac_ch, TAXONOMY_MERGE.out.taxa_mmseqs_ch, CHECKV.out.checkv2vContigs_ch, VIRSORTER2.out.vs2_score_ch, VIBRANT.out.vibrant_quality_ch, DVF.out.dvf2vContigs_ch, ch_replicyc)
 
         CUSTOM_DUMPSOFTWAREVERSIONS (
             ch_versions.unique().collectFile(name: 'collated_versions.yml')