Skip to content

Latest commit

 

History

History
64 lines (46 loc) · 2.55 KB

README.md

File metadata and controls

64 lines (46 loc) · 2.55 KB
Raw

wafergen-scrna

Summary

This pipeline analyses WaferGen’s ICELL8 single cell sequencing data. It uses umis and scRNApipe for the core analysis.

In a first step reads are quality trimmed and adapter sequences are removed (both achieved with Skewer). Reads are then further processed by the umis package (facilitated by scRNApipe). The preprocessed reads are aligned by two different methods

  1. with STAR against the given reference genome and
  2. with Kallisto against the transcripts.

The mapped BAM files from STAR aligner are deduplicated (user option) and meta-features (genes) are counted using featurecounts. The pseudo-aligned reads with transcripts are counted by unique UMI (gene-UMI). The analysis provides the expression matrix of genes and transcripts for cell barcodes.

The pipeline also provides generic stats, coverage, mappability, MultiQC, RNA-SeQC (per cellular barcode) and FastQC output.

Input:

Cellular barcode list (see option -c): All samples barcodes must be listed in this file (for non-demux use "1")

See cfg/references.yaml for reference data (reference genome, index, annotation and gene id files) used by default (also refer to option --references-cfg).

In addition you will need to provide sample definitions, either through a config file (--sample-cfg) or through a full definition on the commandline (options -1, -2 and -s).

Output:

The following lists the most important files/directories that are created per sample in correspondingly named subfolders of out:

  • Results/gene_ExpressionMatrix.csv: Expression matrix from STAR aligned BAM file
  • kallisto/kallisto.tagcount.txt: Expression matrix by Kallisto
  • mb-histogram.txt: Number of of raw reads from each cell barcode-UMI
  • cb-histogram.txt: Number of of raw reads from each cell barcode

Reports:

  • Reports/FinalReport.html: MultiQC output
  • Reports//QCreports/{sample}_fastqc.html: FastQC output
  • Reports/SummarisedStats.csv: Summary stats of reads preprocessing, STAR, featurecount, dedup logs etc.

Notes

Deduplication can be very slow for large data-sets. We recommend to not use it unless necessary (--dedup).

References