Figure 2: Ensembl gene and isoform annotations generally increased between 2014 & 2023, and reveal complex RNA isoform diversity.

For figures (a-d), a representative Ensembl annotation was chosen for each year from 2014–2023. (a) The number of annotated genes per year from 2014–2023. After 2016, the overall trend is an increase, with a reasonably sharp increase around 2020. We see that some annotations appear to be dropped completely between years (Venn diagram). (b) Same as (a) but restricted to protein-coding genes. In contrast, there were more fluctuations in the number of annotated protein-coding genes, though still a positive trend. One gene was annotated as protein-coding in the 2019 and 2023 annotations but was not in 2021. (c) Looking specifically at RNA isoform annotations, the trend is positive, with a sharp increase in 2020. (d) Similar patterns exist for RNA isoforms annotated as protein-coding. There are five isoforms that were not labeled as protein-coding in 2021 but were in 2019 and 2023. (e) Bar plot showing transcript biotype for all isoforms in 2023. As expected, protein-coding was the most common biotype. Interestingly, retained intron was 3^rd and nonsense mediated decay was 5^th. (f) Histogram showing number of annotated isoforms per gene. Colored, zoomed subplots shown for convenience. The majority of gene bodies had only one annotated isoform (median = 1; 75^th percentile: 4; 85^th: 8; 95^th: 16), but some had more than 100. The most annotated isoforms for a single gene body was 296 (PCBP1-AS1; ENSG00000179818). (g) Similar to (f) but showing the number of isoforms per protein-coding gene. The median number of annotated isoforms is 6 (75^th percentile: 11 isoforms). The most annotations for a single protein-coding gene body was 192 (MAPK10; ENSG00000109339).