inst/human_ensembl_to_hgnc_entrez/README.md

In Benjamin-Vincent-Lab/PostRNASeqAlign: Functions for writing commands and post processing the Star Salmon pipeline.

human_ensembl_to_hgnc_entrez v0.1-01 https://sc.unc.edu/benjamin-vincent-lab/gene-sigs/human_ensembl_to_hgnc_entrez

This script was made to make a table that we can use to convert ENST transcripts to HGNC and entrez gene ids. cDNA transcript features have also been added.

Conversions

Using biomaRt and AnnotationDbi is a pretty awful way to go about converting genes. The resources are slow, lose connections, are down a lot, and have tons of gaps in the data. Contrarily, the GTF file has most of everything we need. To get this information we have to scrape it ourselves (GTF extraction tools won't get it :-/). If we combine that with the grch38 cDNA and we have everything to perfectly map the hgnc to the transcripts and pull out cDNA features. Entrez IDs are less complete. Entrez were better for matching to UCSC transcripts a while back, but aren't near as complate as HGNC anymore. At this point it's highly recommended to map reads to ENST and convert to HGNC from there.

cDNA

In addition to doing the conversion of ensembl to different genes this script also gets the cDNA GC content and length. The biomaRt tx_length and percentage_gc_content are for the full length transcript including introns. cds_length should be the length of the coding sequence (sum of exon lengths), but it's not even clsoe to the cDNA length. The full length cDNA is not included in the final matrix to reduce the size of the table that will go into the R package but can be found in the cDNA intermdiate file.

Type and description

Information on the types (biomaRt name_1006) and description are provided for look-up separated by pipes so they can be easily searched for piped terms (eg '|AB|' will not show up in '|ABAB|')

UCSC caveats

USCS transcript ids, while on the table, were not the principal focus here so not a lot of effort was made to make those mappings 100% complete. Plus adding UCSC would have duplicated a lot of ENST rows and made mapping to our principal goal less clear.

Benjamin-Vincent-Lab/PostRNASeqAlign documentation built on Feb. 14, 2024, 8:31 a.m.