feat(rules): support CellRanger

Due to ongoing file formatting issues with STAR, we have reverted to Cellranger. This commits also add support for the necessary mkgtf and mkref calls required for a custom cellranger reference.

BREAKING CHANGE: Though the reversion does not impact apparent function significantly, the outputs and underlying analysis are changed enough to warrant this BREAKING.

workflow/rules/cellranger.smk

@@ -0,0 +1,92 @@
+rule mkgtf:
+    input:
+        gtf=rules.move_coordinates.output.gtf,
+        bin=rules.get_cellranger.output.bin,
+    output:
+        gtf="resources/genome_filtered.gtf",
+    log:
+        "results/logs/mkgtf/mkgtf.log",
+    benchmark:
+        "results/benchmarks/mkgtf/mkgtf.txt"
+    shell:
+        "{input.bin} "
+        "mkgtf "
+        "{input.gtf} "
+        "{output.gtf} "
+        "--attribute=gene_biotype:protein_coding "
+        "--attribute=gene_biotype:lincRNA "
+        "--attribute=gene_biotype:antisense "
+        "--attribute=gene_biotype:IG_LV_gene "
+        "--attribute=gene_biotype:IG_V_gene "
+        "--attribute=gene_biotype:IG_V_pseudogene "
+        "--attribute=gene_biotype:IG_D_gene "
+        "--attribute=gene_biotype:IG_J_gene "
+        "--attribute=gene_biotype:IG_J_pseudogene "
+        "--attribute=gene_biotype:IG_C_gene "
+        "--attribute=gene_biotype:IG_C_pseudogene "
+        "--attribute=gene_biotype:TR_V_gene "
+        "--attribute=gene_biotype:TR_V_pseudogene "
+        "--attribute=gene_biotype:TR_D_gene "
+        "--attribute=gene_biotype:TR_J_gene "
+        "--attribute=gene_biotype:TR_J_pseudogene "
+        "--attribute=gene_biotype:TR_C_gene"
+
+
+rule mkref:
+    input:
+        gtf=rules.mkgtf.output.gtf,
+        fa=rules.get_fa.output.fa,
+        bin=rules.get_cellranger.output.bin,
+    output:
+        ref=directory("resources/ref_genome"),
+    log:
+        "results/logs/mkref/mkref.log",
+    benchmark:
+        "results/benchmarks/mkref/mkref.txt"
+    shell:
+        "{input.bin} "
+        "mkref "
+        "--genome={output.ref} "
+        "--genes={input.gtf} "
+        "--fasta={input.fa} "
+
+
+rule count:
+    input:
+        R1="data/{sample}_S1_L00{lane}_R1_001.fastq.gz",
+        R2="data/{sample}_S1_L00{lane}_R2_001.fastq.gz",
+        bin=rules.get_cellranger.output.bin,
+        genome=rules.mkref.output.ref,
+    output:
+        results=directory("results/counts/{sample}_{lane}/outs/raw_feature_bc_matrix"),
+        mtx="results/counts/{sample}_{lane}/outs/raw_feature_bc_matrix/matrix.mtx.gz",
+        html=report(
+            "results/counts/{sample}_{lane}/outs/web_summary.html",
+            caption="../reports/counts.rst",
+            category="2. Cellranger Counts",
+            subcategory="{sample}_{lane}",
+        ),
+    params:
+        introns=convert_introns(),
+        n_cells=config["counts"]["n_cells"],
+        mem=config["counts"]["mem"],
+    log:
+        "results/logs/counts/{sample}_{lane}.log",
+    benchmark:
+        "results/benchmarks/counts/{sample}_{lane}.txt"
+    threads: 16
+    shell:
+        "{input.bin} "
+        "count "
+        "--nosecondary "
+        "{params.introns} "
+        "--id {wildcards.sample}_{wildcards.lane} "
+        "--transcriptome {input.genome}  "
+        "--fastqs data "
+        "--sample {wildcards.sample} "
+        "--lanes {wildcards.lane} "
+        "--expect-cells {params.n_cells} "
+        "--localcores {threads} "
+        "--localmem {params.mem} "
+        "&> {log} && "
+        "mv {wildcards.sample}_{wildcards.lane} results/counts/{wildcards.sample}_{wildcards.lane} "

workflow/rules/star_resources.smk → workflow/rules/cellranger_resources.smk

@@ -1,21 +1,21 @@
 # Gather input resources for STAR
 
 
-rule get_whitelist:
-    params:
-        url=config["get_whitelist"]["url"],
+rule get_cellranger:
     output:
-        wl="resources/whitelist.txt.gz",
+        cr=directory("resources/cellranger"),
+        bin="resources/cellranger/bin/cellranger",
+    params:
+        url=config["get_cellranger"]["url"],
     log:
-        "results/logs/get_whitelist/get_whitelist.log",
+        "results/logs/get_cellranger/get_cellranger.log",
     benchmark:
-        "results/benchmarks/get_whitelist/get_whitelist.txt"
+        "results/benchmarks/get_cellranger/get_cellranger.txt"
     shell:
-        "wget "
-        "--no-verbose "
-        "-O {output.wl} "
-        "{params.url} "
-        "&> {log}"
+        "wget --no-verbose -O- {params.url} | "
+        "tar -xzf - -C resources && "
+        "rm -rf resources/cellranger.tar.gz && "
+        "mv resources/cellranger-* resources/cellranger "
 
 
 rule get_gtf: