This directory contains gene and biotype assignments of all "non-anchored" compmerge non-redudnant transcript models (TMs) in human and mouse. 

They were obtained by comparing the genomic coordinates of TMs and GENCODE 20 (human) to GENCODE M3 (mouse) annotations, plus extra, non-GENCODE probed features.

# File naming scheme:

<species>.compmergeId.To.GENCODEgene_id.To.biotype.tsv

where:
	species:
		"mm": mouse
		"hs": human

# File format (tab-separated):

There is one line per TM.

    column 1: TM compmerge identifier. Contains the tissue in which the model was built (when set to "pooled", the merging was made across all captured tissues).
    column 2: comma-separated list of "annotated gene / annotated biotype values", that the corresponding TM overlaps, in the form:

				<gene_id#1>:<gene_id#1's annotated biotype>,<gene_id#2>:<gene_id#2's annotated biotype>,[...]


# Note about "biotype" values:

- The following GENCODE gene types were tagged “lncRNA”: “antisense”, "lincRNA", "processed_transcript", "sense_intronic" and "sense_overlapping".
- In our analysis, we consider any read overlapping multiple genes of distinct biotypes as "multi-biotype"