TEquant

Tools for manipulating quantification of transposable element reads from RNA-seq data.

Workflow

• Map reads to reference

• Create a GTF

  • Curated TE gtf files are available from TEToolkit, here: http://labshare.cshl.edu/shares/mhammelllab/www-data/TEToolkit/ (github page: https://github.com/mhammell-laboratory/tetoolkit)
  • Can combine transcript and repeatmasker GTF files, filtering out repeats that lie in annotated transcripts

• Create a table relating different levels of the TE heirarchy

  • Sample code for this (that works for the hg19_rmsk_TE.gtf file) is in parse_TE_heirarchies.py.

• Run featurecounts on each sample

  • Settings depend on library type (e.g. stranded, unstranded)
  • Can ignore or keep multimapping reads
  • Note that featurecounts will by default throw away reads that overlap multiple features by default

• Merge featurecounts outputs across samples

  • Using merge_featurecounts_lowmem.py
  • This generates a file with counts across all samples, and some metadata (e.g. total number of mapped reads in each sample), indicated by a '_' prefix (e.g. '_assigned' for reads assigned to a feature - see featurecounts documentation for details).

• Subselect and sum data across transcript types:

  • Using split_and_sum_TE_counts
  • Creates a count matrix for each type of transcript:
    • gene_id
    • TEclass_id (most general)
    • TEfamily_id
    • TEgene_id
    • TEtranscript_id (most specific - individual loci)