Merge branch 'cellranger-4.0.0_ref-2020A' into '50-seurat.object'

Cellranger 4.0.0 ref 2020 a

See merge request !75

workflow/conf/aws.config

-workDir = 's3://'
-aws.client.storageEncryption = 'AES256'
-aws {
-  region = ''
-  batch {
-    cliPath = '/home/ec2-user/miniconda/bin/aws'
-  }
-}
-
-process {
-  executor = 'awsbatch'
-  queue = 'default-'
-  cpus = 1
-  memory = '1 GB'
-
-  withLabel: checkDesignFile {
-    cpus = 2
-    memory = '1 GB'
-  }
-  withLabel: count211 {
-    cpus = 2
-    memory = '30 GB'
-  }
-  withLabel: count220 {
-    cpus = 2
-    memory = '30 GB'
-  }
-  withLabel: count302 {
-    cpus = 2
-    memory = '30 GB'
-  }
-  withLabel: count310 {
-    cpus = 2
-    memory = '30 GB'
-  }
-  withLabel: count400 {
-    cpus = 2
-    memory = '30 GB'
-  }
-  withLabel: versions {
-    cpus = 3
-    memory = '1 GB'
-  }
-  withLabel: multiqc {
-    cpus = 1
-    memory = '1 GB'
-  }
-}

workflow/conf/biohpc.config

-params {
-  // Reference file paths on BioHPC
-  genomes {
-    'GRCh38-2020-A' {
-      loc = '/project/apps_database/cellranger/refdata-gex-'
-    }
-    'GRCh38-3.0.0' {
-      loc = '/project/apps_database/cellranger/refdata-cellranger-'
-    }
-    'GRCh38-1.2.0' {
-      loc = '/project/apps_database/cellranger/refdata-cellranger-'
-    }
-    'hg19-3.0.0' {
-      loc = '/project/apps_database/cellranger/refdata-cellranger-'
-    }
-    'hg19-1.2.0' {
-      loc = '/project/apps_database/cellranger/refdata-cellranger-'
-    }
-    'mm10-2020-A' {
-      loc = '/project/apps_database/cellranger/refdata-gex-'
-    }
-    'mm10-3.0.0' {
-      loc = '/project/apps_database/cellranger/refdata-cellranger-'
-    }
-    'mm10-1.2.0' {
-      loc = '/project/apps_database/cellranger/refdata-cellranger-'
-    }
-    'GRCh38_and_mm10-2020-A' {
-      loc = '/project/apps_database/cellranger/refdata-gex-'
-    }
-    'GRCh38_and_mm10-3.1.0' {
-      loc = '/project/apps_database/cellranger/refdata-cellranger-'
-    }
-    'hg19_and_mm10-3.0.0' {
-      loc = '/project/apps_database/cellranger/refdata-cellranger-'
-    }
-    'hg19_and_mm10-1.2.0' {
-      loc = '/project/apps_database/cellranger/refdata-cellranger-'
-    }
-    'ercc92-1.2.0' {
-      loc = '/project/apps_database/cellranger/refdata-cellranger-'
-    }
-  }
-  // Chemistry mapping parameter
-  chemistry {
-    'auto' {
-      param = 'auto'
-    }
-    '3GEXv1' {
-      param = 'SC3Pv1'
-    }
-   '3GEXv2' {
-      param = 'SC3Pv2'
-    }
-   '3GEXv3' {
-      param = 'SC3Pv3'
-    }
-   '5GEX' {
-      param = 'fiveprime'
-    }
-  }
-}
-
-singularity {
-  enabled = true
-  cacheDir = '/project/BICF/BICF_Core/shared/gudmap/singularity_cache/'
-}
-
-env {
-  http_proxy = 'http://proxy.swmed.edu:3128'
-  https_proxy = 'http://proxy.swmed.edu:3128'
-  all_proxy = 'http://proxy.swmed.edu:3128'
-}

workflow/conf/cluster.config

-process {
-  executor = 'slurm'
-  queue = '32GB'
-  clusterOptions = '--hold'
-
-  withLabel: trackStart {
-    executor = 'local'
-  }
-  withLabel: checkDesignFile {
-    executor = 'local'
-  }
-  withLabel: count211 {
-    queue = '128GB,256GB,256GBv1,384GB'
-  }
-  withLabel: count220 {
-    queue = '128GB,256GB,256GBv1,384GB'
-  }
-  withLabel: count302 {
-    queue = '128GB,256GB,256GBv1,384GB'
-  }
-  withLabel: count310 {
-    queue = '128GB,256GB,256GBv1,384GB'
-  }
-  withLabel: count400 {
-    queue = '128GB,256GB,256GBv1,384GB'
-  }
-  withLabel: versions {
-    executor = 'local'
-  }
-  withLabel: multiqc {
-    executor = 'local'
-  }
-}

workflow/conf/local.config

-process {
-  executor = 'local'
-}
\ No newline at end of file

workflow/conf/multiqc_config.yaml

-# Custom Logo
-custom_logo: 'bicf_logo.png'
-custom_logo_url: 'https://www.utsouthwestern.edu/labs/bioinformatics/'
-custom_logo_title: 'Bioinformatics Core Facility'
-
-report_header_info:
-    - Contact E-mail: 'bicf@utsouthwestern.edu'
-    - Application Type: 'cellranger_count'
-    - Department: 'Bioinformatic Core Facility, Department of Bioinformatics'
-
-
-# Title to use for the report.
-title: BICF CellRanger Count Analysis Report
-
-report_comment: >
-  This report has been generated by the <a href="https://git.biohpc.swmed.edu/BICF/Astrocyte/cellranger_count"
-  target="_blank">BICF/cellranger_count</a> pipeline.
-
-custom_data:
-  metrics_summary:
-    file_format: 'tsv'
-    id: 'metrics_summary'
-    contents: 'Estimated Number of Cells        Mean Reads per Cell     Median Genes per Cell   Number of Reads Valid Barcodes  Sequencing Saturation   Q30 Bases in Barcode    Q30 Bases in RNA Read   Q30 Bases in UMI        Reads Mapped to Genome  Reads Mapped Confidently to Genome      Reads Mapped Confidently to Intergenic Regions  Reads Mapped Confidently to Intronic Regions    Reads Mapped Confidently to Exonic Regions      Reads Mapped Confidently to Transcriptome       Reads Mapped Antisense to Gene  Fraction Reads in Cells Total Genes Detected    Median UMI Counts per Cell'
-    section_name: 'Metrics Summary'
-    plot_type: 'generalstats'
-
-sp:
-  metrics_summary:
-    fn: 'metrics_summary_mqc.tsv'
-
-table_columns_placement:
-  metrics_summary:
-    Estimated Number of Cells: 1
-    Mean Reads per Cell: 2
-    Median Genes per Cell: 3
-    Number of Reads: 4
-    Sequencing Saturation: 5
-    Reads Mapped Confidently to Genome: 6
-    Reads Mapped Confidently to Transcriptome: 7
-    Fraction Reads in Cells: 8
-    Total Genes Detected: 9
-    Median UMI Counts per Cell: 10
-    Valid Barcodes: 1100
-    Reads Mapped Antisense to Gene: 1200
-
-table_columns_visible:
-  metrics_summary:
-    Q30 Bases in Barcode: False
-    Q30 Bases in RNA Read: False
-    Q30 Bases in UMI: False
-    Reads Mapped to Genome: False
-    Reads Mapped Confidently to Intergenic Regions: False
-    Reads Mapped Confidently to Intronic Regions: False
-    Reads Mapped Confidently to Exonic Regions: False
-
-thousandsSep_format: ''
-
-report_section_order:
-    software_versions:
-      order: -1100
-    software_references:
-      order: -1200

workflow/configs/aws.config

@@ -12,27 +12,27 @@ process {
   cpus = 1
   memory = '1 GB'
 
-  withLabel: checkDesignFile {
+  withName: checkDesignFile {
     cpus = 2
     memory = '1 GB'
   }
-  withLabel: count211 {
+  withName: count211 {
     cpus = 2
     memory = '30 GB'
   }
-  withLabel: count302 {
+  withName: count302 {
     cpus = 2
     memory = '30 GB'
   }
-  withLabel: count310 {
+  withName: count310 {
     cpus = 2
     memory = '30 GB'
   }
-  withLabel: versions {
+  withName: versions {
     cpus = 3
     memory = '1 GB'
   }
-  withLabel: multiqc {
+  withName: multiqc {
     cpus = 1
     memory = '1 GB'
   }

workflow/configs/biohpc.config

 params {
   // Reference file paths on BioHPC
   genomes {
+    'GRCh38-2020-A' {
+      loc = '/project/apps_database/cellranger/refdata-gex-'
+    }
     'GRCh38-3.0.0' {
       loc = '/project/apps_database/cellranger/refdata-cellranger-'
     }
@@ -13,12 +16,18 @@ params {
     'hg19-1.2.0' {
       loc = '/project/apps_database/cellranger/refdata-cellranger-'
     }
+    'mm10-2020-A' {
+      loc = '/project/apps_database/cellranger/refdata-gex-'
+    }
     'mm10-3.0.0' {
       loc = '/project/apps_database/cellranger/refdata-cellranger-'
     }
     'mm10-1.2.0' {
       loc = '/project/apps_database/cellranger/refdata-cellranger-'
     }
+    'GRCh38_and_mm10-2020-A' {
+      loc = '/project/apps_database/cellranger/refdata-gex-'
+    }
     'GRCh38_and_mm10-3.1.0' {
       loc = '/project/apps_database/cellranger/refdata-cellranger-'
     }
@@ -52,6 +61,11 @@ params {
   }
 }
 
+singularity {
+  enabled = true
+  cacheDir = '/project/BICF/BICF_Core/shared/gudmap/singularity_cache/'
+}
+
 env {
   http_proxy = 'http://proxy.swmed.edu:3128'
   https_proxy = 'http://proxy.swmed.edu:3128'

workflow/configs/cluster.config

 process {
   executor = 'slurm'
-  queue = 'super'
+  queue = '32GB'
   clusterOptions = '--hold'
 
-  withName:trackStart {
+  withName: trackStart {
     executor = 'local'
   }
-  withName:checkDesignFile {
+  withName: checkDesignFile {
     executor = 'local'
   }
-  withName:count211 {
+  withName: count211 {
     queue = '128GB,256GB,256GBv1,384GB'
   }
-  withName:count302 {
+  withName: count220 {
     queue = '128GB,256GB,256GBv1,384GB'
   }
-  withName:count310 {
+  withName: count302 {
     queue = '128GB,256GB,256GBv1,384GB'
   }
-  withName:versions {
+  withName: count310 {
+    queue = '128GB,256GB,256GBv1,384GB'
+  }
+  withName: count400 {
+    queue = '128GB,256GB,256GBv1,384GB'
+  }
+  withName: versions {
     executor = 'local'
   }
-  withName:multiqc {
+  withName: multiqc {
     executor = 'local'
   }
 }

workflow/main.nf

@@ -22,7 +22,6 @@ params.name = "run"
 params.fastq = "test_data/mu.v3s1r500/*.fastq.gz"
 params.designFile = "test_data/mu.v3s1r500/design.csv"
 params.genome = 'mm10-2020-A'
-params.genomeLocation = '/project/apps_database/cellranger/refdata-gex-'
 params.expectCells = 10000
 params.forceCells = 0
 params.kitVersion = '3GEXv3'
@@ -39,7 +38,11 @@ if (params.kitVersion == "3GEXv3" && params.version == '2.1.1') {
 // Define variables if astrocyte (or from config)
 if (params.astrocyte) {
   print("Running under astrocyte")
-  params.genomeLocation = '/project/apps_database/cellranger/refdata-cellranger-'
+  if (params.version == "4.0.0") {
+    params.genomeLocation = '/project/apps_database/cellranger/refdata-gex-'
+  } else {
+    params.genomeLocation = '/project/apps_database/cellranger/refdata-cellranger-'
+  }
   if (params.kitVersion == "3GEXv1") {
     params.chemistryParam ='SC3Pv1'
   } else if (params.kitVersion == "3GEXv2") {
@@ -169,7 +172,7 @@ refLocation.into {
 expectCells211 = expectCells
 expectCells302 = expectCells
 expectCells310 = expectCells
-expectCells3400 = expectCells
+expectCells400 = expectCells
 forceCells211 = forceCells
 forceCells302 = forceCells
 forceCells310 = forceCells
@@ -330,7 +333,7 @@ process count400 {
 
   input:
     set sample, file("${sample}_S?_L001_R1_001.fastq.gz"), file("${sample}_S?_L001_R2_001.fastq.gz"), file(script) from samples400
-    file ref from refLocation310.first()
+    file ref from refLocation400.first()
     expectCells400
     forceCells400
     chemistryParam400

workflow/nextflow.config

 profiles {
   standard {
     includeConfig 'configs/biohpc.config'
+    includeConfig 'configs/cluster.config'
   }
   biohpc {
     includeConfig 'configs/biohpc.config'