Merge pull request #14 from sanger-tol/raw_asm

Add raw assembly stage
sanger-tol · Jun 1, 2023 · 275d7b7 · 275d7b7
2 parents bbf9725 + d9e98c3
commit 275d7b7
Show file tree

Hide file tree

Showing 28 changed files with 996 additions and 118 deletions.
diff --git a/assets/test.yaml b/assets/test.yaml
@@ -1,17 +1,13 @@
 samples:
-  - id: ilEupCent1
-    assembly:
-      primary: /lustre/scratch124/tol/projects/darwin/users/kk16/development/nextflow/sanger-tol-genomeassembly/test/primary.fa
-      haplotigs: /lustre/scratch124/tol/projects/darwin/users/kk16/development/nextflow/sanger-tol-genomeassembly/test/haplotigs.fa
+  - id: baUndUnlc1
     illumina_10X:
-      reads: /lustre/scratch124/tol/projects/darwin/users/kk16/development/nextflow/sanger-tol-genomeassembly/test/10X/small/
+      reads: /lustre/scratch123/tol/resources/nextflow/test-data/Undibacterium_unclassified/genomic_data/baUndUnlc1/10x/
     pacbio:
       reads:
-        - reads: /lustre/scratch124/tol/projects/darwin/users/kk16/development/nextflow/sanger-tol-genomeassembly/test/1000_reads.fasta
+        - reads: /lustre/scratch123/tol/resources/nextflow/test-data/Undibacterium_unclassified/genomic_data/baUndUnlc1/pacbio/fasta/HiFi.reads.fasta
     HiC:
       reads:
-        - reads: /lustre/scratch124/tol/projects/darwin/users/kk16/development/nextflow/sanger-tol-genomeassembly/test/hic.01.cram
+        - reads: /lustre/scratch123/tol/resources/nextflow/test-data/Undibacterium_unclassified/genomic_data/baUndUnlc1/hic-arima2/41741_2#7.sub.cram
       arima_motif: GATC,GANTC,CTNAG,TTAA
     busco:
-      lineages_path: /lustre/scratch123/tol/resources/busco/v5/
-      lineage: insecta_odb10
+      lineage: bacteria_odb10
diff --git a/assets/test_gsMetZobe1.yaml b/assets/test_gsMetZobe1.yaml
@@ -0,0 +1,14 @@
+samples:
+  - id: gsMetZobe1
+    illumina_10X:
+      reads: /lustre/scratch124/tol/projects/darwin/users/kk16/development/nextflow/gsMetZobe1/10x/
+    pacbio:
+      reads:
+        - reads: /lustre/scratch124/tol/projects/darwin/users/kk16/development/nextflow/gsMetZobe1/pacbio/m64125_200823_145825.ccs.bc1019_BAK8B_OA--bc1019_BAK8B_OA.filtered.fasta.gz
+    HiC:
+      reads:
+        - reads: /lustre/scratch124/tol/projects/darwin/users/kk16/development/nextflow/gsMetZobe1/hic-arima/35528_4#7.cram
+      arima_motif: GATC,GANTC,CTNAG,TTAA
+    busco:
+      lineages_path: /lustre/scratch123/tol/resources/busco/v5/
+      lineage: fungi_odb10
diff --git a/assets/test_iyVesGerm1.yaml b/assets/test_iyVesGerm1.yaml
@@ -4,11 +4,10 @@ samples:
       primary: /lustre/scratch124/tol/projects/darwin/data/insects/Vespula_germanica/working/iyVesGerm1.hifiasm.20201015/iyVesGerm1.p_ctg.fa.gz
       haplotigs: /lustre/scratch124/tol/projects/darwin/data/insects/Vespula_germanica/working/iyVesGerm1.hifiasm.20201015/iyVesGerm1.a_ctg.fa.gz
     illumina_10X:
-      reads: /lustre/scratch124/tol/projects/darwin/data/insects/Vespula_germanica/genomic_data/iyVesGerm1/10x/
+      reads: /lustre/scratch124/tol/projects/darwin/users/kk16/development/nextflow/sanger-tol-genomeassembly/test/10X_iyVesGerm1
     pacbio:
       reads:
         - reads: /lustre/scratch124/tol/projects/darwin/data/insects/Vespula_germanica/genomic_data/iyVesGerm1/pacbio/fasta/m64094_200217_145414.ccs.bc1011_BAK8A_OA--bc1011_BAK8A_OA.filtered.fasta.gz
-      kmer_pref: /lustre/scratch124/tol/projects/darwin/data/insects/Vespula_germanica/genomic_data/iyVesGerm1/pacbio/kmer/k31/
     HiC:
       reads:
         - reads: /lustre/scratch124/tol/projects/darwin/data/insects/Vespula_germanica/genomic_data/iyVesGerm1/hic-arima2/34957_3#2.cram

diff --git a/conf/base.config b/conf/base.config
@@ -24,6 +24,11 @@ process {
     //        If possible, it would be nice to keep the same label naming convention when
     //        adding in your local modules too.
     // See https://www.nextflow.io/docs/latest/config.html#config-process-selectors
+    withName:HIFIASM {
+        cpus   = { check_max( 28    * task.attempt, 'cpus'    ) }
+        memory = { check_max( 200.GB * task.attempt, 'memory'  ) }
+        time   = { check_max( 48.h  * task.attempt, 'time'    ) }
+    }
     withLabel:process_single {
         cpus   = { check_max( 1                  , 'cpus'    ) }
         memory = { check_max( 6.GB * task.attempt, 'memory'  ) }