processTerminalExons: Process predicted terminal exons
In ldg21/SGSeq: Splice event prediction and quantification from RNA-seq data

Description Usage Arguments Details Value Author(s) Examples

Predicted terminal exons are processed as described under Details.

1	processTerminalExons(features, min_overhang = NA)

`features`	`TxFeatures` object
`min_overhang`	Minimum overhang required to suppress filtering or trimming of predicted terminal exons (see Details). Use `NA` to exclude all terminal exons sharing a splice with an internal exon and trim all remaining terminal exons overlapping other exons.

Processing of terminal exon predictions is done in two steps: (1) terminal exons that share a splice site with an internal exon are filtered, and (2) remaining terminal exons that overlap other exons are trimmed.

predictTxFeatures predicts flanking terminal exons for each identified splice junction. This ensures that each splice junction has a flanking exon after merging with mergeTxFeatures. This approach results in many predicted terminal exons that share a splice site with predicted internal exons (often contained within them or with a short overhang due to incorrect alignments). Most of these are not real terminal exons and are filtered before further analysis. Filtering based on the overhang is controlled with argument min_overhang.

Some of the remaining predicted terminal exons overlap other exons such that their unspliced boundary shows a short overhang with respect to a spliced boundary of the overlapping exon. Often these exon extensions into an intron are due to incorrect alignments. Terminal exons with overhang smaller than min_overhang are trimmed such that their trimmmed unspliced boundary coincides with the spliced boundary of the overlapping exon.

TxFeatures object with processed features

Leonard Goldstein