EDGER_EXON fails if only single entry in contrast sheet

[dkoppstein]

Description of the bug

My contrast sheet looks like this:

contrast,treatment,control
KO_CONTROL,KO,CONTROL

then I get

  Loading required package: limma
  Error in asplit(contrasts, MARGIN = 2) : 
    dim(x) must have a positive length

however, making a new fake contrast sheet with also the reverse contrast fixes it.

contrast,treatment,control
KO_CONTROL,KO,CONTROL
CONTROL_KO,CONTROL,KO

The exact place the error occurs is in the script run_edger_exon.R

DGELRT.exprs <- mapply(
    FUN      = glmQLFTest,
    contrast = asplit(contrasts, MARGIN = 2),
    MoreArgs = list(glmfit = DGEGLM),
    SIMPLIFY = FALSE
)

Command used and terminal output

nextflow run \
	/omics/groups/xxxxx/internal/xxxxx/src/github.com/dkoppstein/rnasplice \
	--input config/samplesheet_splicing.csv \
	--contrasts config/contrastsheet_splicing.csv \
	--outdir workspace/rnasplice_results_ko_vs_control \
	-c config/dkfz.config \
	--fasta $(GENOMEDIR)/GRCh38.primary_assembly.genome.fa \
	--gtf $(GENOMEDIR)/gencode.v43.annotation.gtf \
	--star_index $(GENOMEDIR)/genome/index/star \
	--salmon_index $(GENOMEDIR)/genome/index/salmon \
	--gencode \
	--save_reference \
	--min_samps_gene_expr 0 \
	--min_samps_feature_expr 0 \
	--min_samps_feature_prop 0 \
	--min_feature_expr 0 \
	--min_feature_prop 0 \
	--min_gene_expr 0 \
	--miso_genes "ENSG00000004961.15, ENSG00000005302.19, ENSG00000147403.18" \
	-resume
	touch $@

Relevant files

No response

System information

Nextflow version 23.04.1 build 5866
HPC
lsf executor
Singularity container
CentOS Linux
dev

[jma1991]

Thanks David - fixed in #82