Skip to content

reportingNf.md

Latest commit

 

History

History
43 lines (28 loc) · 2.9 KB

reportingNf.md

File metadata and controls

43 lines (28 loc) · 2.9 KB

Reporting-only Workflow

A separate sub-workflow, reporting, is included to run only the QC and reporting steps of the workflow, after alignment and quantification (i.e. after STARsolo). This includes

  • Cell filtering (UMI threshold)
  • Sample metric and report generation
  • Library metric and report generation

This sub-workflow can be used to regenerate outputs after adjusting parameters, without re-running alignment. It is also used to combine (merge) results from multiple different extended throughput plates (libraries), sequenced on different sequencing runs; see extendedThroughput.

Usage

The reporting sub-workflow can be started by running

nextflow run /PATH/TO/ScaleRna -profile ... --genome ... --samples samples.csv --reporting --resultDir <outDir from previous workflow run>

where samples.csv and genome should be the same files as used in the previous (alignment) run of the workflow.

Inputs

The reporting workflow will read the (raw) STARSolo output and a STARSolo log file from the previous pipeline run specified in resultDir. Specifically, the raw output is read from resultDir/alignment/<sample>.<libName>.star.solo and the log file is read from resultDir/alignment/<sample>.<libName>.star.align It also reads reference information from the genome.json and library.json

Outputs

The reporting workflow produces

  • A new filtered gene expression matrix (samples/<sample>.<libName>.filtered/)
  • Cell metrics (samples/<sample>.<libName>.allCells.csv)
  • New per-sample QC reports (reports/<sample>/<sample>.<libName>.report.html)
    • And .csv metric files

Currently the reporting sub-workflow does not aggregate read-trimming statistics from the original run, so the Total Sample Reads (pre-trimming) and Average Trimmed Read Length metrics will be missing from the re-generated reports.

Combining Analysis Runs

The reporting sub-workflow can be used to combine results from multiple analysis workflow runs. To do this, multiple paths to previous workflow outputs can be specified in a resultDir column in samples.csv, instead of the --resultDir command-line option:

sample resultDir
pbmc1 /PATH/TO/RUN1/ScaleRna.out
pbmc2 /PATH/TO/RUN2/ScaleRna.out

See samples.ext-merge.csv for a complete example combining cells from multiple extended throughput plates. If --merge is set (default), the workflow produces combined outputs for each sample across all libraries (plates).

Note that this function only combines cells or samples from different runs, not reads for the same set of cells. Hence it cannot be used to combine two sequencing runs of the same library (multiple sets of reads for the same cells). Those need to be combined at the fastq level and go through alignment together in order to detect duplicate reads across fastq files.