detect_alignment_errors: Detect the most common type of alignment errors in...

In Maxim-Ivanov/TranscriptomeReconstructoR: Data-driven Annotation of the Transcriptome

Detect the most common type of alignment errors in full-length RNA-seq reads aligned by Minimap2

Description

Detect the most common type of alignment errors in full-length RNA-seq reads aligned by Minimap2

Usage

detect_alignment_errors(long_reads, min_read_support = 2)

Arguments

long_reads	GRangesList object.
min_read_support	Positive integer.

Value

GRangesList object which is identical to the input object, except for the extra logical column aln_error.

Details

The input GRangesList object is returned by the adjust_exons_of_long_reads() function.
The Minimap2 aligner is prone to under-split the full-length RNA-seq reads. In certain reads, two adjacent exons may appear as a single subalignment. Assuming that most of the reads are still aligned correctly, one can detect alignment errors by comparing each subalignment to the set of constitutive exons. The constitutive exons are defined here as exonic intervals supported by more than 50% of reads in given locus (but not less than min_read_support reads). Subalignment with an alternative 5'- and/or 3-border (relative to a constitutive exon) are considered valid alternative exons, only if the next constitutive exon is also present in the read. Otherwise, the alignment is marked as a possible alignment error.

Maxim-Ivanov/TranscriptomeReconstructoR documentation built on Oct. 3, 2023, 11:19 p.m.