Add --tax_evalue parameter

Addresses #56
Golob-Minot · Jan 28, 2021 · 1deff83 · 1deff83
1 parent 17d0f10
commit 1deff83
Show file tree

Hide file tree

Showing 3 changed files with 9 additions and 1 deletion.
diff --git a/main.nf b/main.nf
@@ -55,6 +55,7 @@ params.famli_batchsize = 10000000 // FAMLI
 // Annotation options
 params.noannot = false
 params.taxonomic_dmnd = false
+params.tax_evalue = 0.00001
 params.ncbi_taxdump = "ftp://ftp.ncbi.nlm.nih.gov/pub/taxonomy/taxdump.tar.gz"
 params.eggnog_db = false
 params.eggnog_dmnd = false
@@ -113,6 +114,7 @@ def helpMessage() {
                             Individual annotations can also be disabled by, e.g., setting --eggnog_db false
       --taxonomic_dmnd      Database used for taxonomic annotation (default: false)
                             (Data available at s3://fh-ctr-public-reference-data/tool_specific_data/geneshot/2020-01-15-geneshot/DB.refseq.tax.dmnd)
+      --tax_evalue          Maximum E-value threshold for taxonomic annotation by DIAMOND alignment
       --ncbi_taxdump        Reference describing the NCBI Taxonomy
                             (default: ftp://ftp.ncbi.nlm.nih.gov/pub/taxonomy/taxdump.tar.gz)
       --eggnog_dmnd         One of two databases used for functional annotation with eggNOG (default: false)
@@ -225,6 +227,7 @@ include {
     eggnog_dmnd: params.eggnog_dmnd,
     taxonomic_dmnd: params.taxonomic_dmnd,
     gencode: params.gencode,
+    tax_evalue: params.tax_evalue,
 )
 
 // Import the workflows used for alignment-based analysis

diff --git a/modules/assembly.nf b/modules/assembly.nf
@@ -4,6 +4,7 @@ container__experiment_collection = "quay.io/fhcrc-microbiome/experiment-collecti
 
 // Default parameters
 params.output_prefix = "geneshot"
+params.tax_evalue = 0.00001
 
 // Containers
 container__assembler = "quay.io/biocontainers/megahit:1.2.9--h8b12597_0"
@@ -515,7 +516,8 @@ diamond \
     --out genes.tax.aln.gz \
     --outfmt 102 \
     --id ${params.min_identity} \
-    --top ${100 - params.min_identity} \
+    --top 5 \
+    --evalue ${params.tax_evalue} \
     --block-size ${task.memory.toMega() / (1024 * 6)} \
     --threads ${task.cpus} \
     --compress 1

diff --git a/run_annotations.nf b/run_annotations.nf
@@ -20,6 +20,7 @@ params.help = false
 params.min_coverage = 50
 params.min_identity = 90
 params.taxonomic_dmnd = false
+params.tax_evalue = 0.00001
 params.ncbi_taxdump = "ftp://ftp.ncbi.nlm.nih.gov/pub/taxonomy/taxdump.tar.gz"
 params.eggnog_db = false
 params.eggnog_dmnd = false
@@ -44,6 +45,7 @@ def helpMessage() {
       --output_hdf          Name of output HDF to be placed in the output folder
       --taxonomic_dmnd      Database used for taxonomic annotation (default: false)
                             (Data available at s3://fh-ctr-public-reference-data/tool_specific_data/geneshot/2020-01-15-geneshot/DB.refseq.tax.dmnd)
+      --tax_evalue          Maximum E-value threshold for taxonomic annotation by DIAMOND alignment
       --ncbi_taxdump        Reference describing the NCBI Taxonomy
                             (default: ftp://ftp.ncbi.nlm.nih.gov/pub/taxonomy/taxdump.tar.gz)
       --eggnog_dmnd         One of two databases used for functional annotation with eggNOG (default: false)
@@ -90,6 +92,7 @@ include { annotation_wf } from './modules/assembly' params(
     eggnog_db: params.eggnog_db,
     eggnog_dmnd: params.eggnog_dmnd,
     taxonomic_dmnd: params.taxonomic_dmnd,
+    tax_evalue: params.tax_evalue,
     noannot: params.noannot
 )