Our lncRNA gene collection is a composite of three previously published lncRNA sources – Gencode [4], the Broad Institute [67], and our own [5] – and three additional public lncRNA collections (see Methods). The methods used in the compilation of these lncRNA collections rely on a combination of full-length cDNAs, deep-coverage RNAseq, targeted RTPCR validation, and extensive manual curation. Therefore, the lncRNA genes that we used are largely as reliable in terms of their underlying evidence and annotation quality as protein-coding genes, and the differences we have uncovered relative to their protein-coding counterparts are not likely to be due to annotation disparities.
