PAN-GO annotations for cardiac muscle contraction

[thomaspd]

From @doughowe
A user has contacted ZFIN with a question about cross-species annotation among mouse, human, and zebrafish for the term "cardiac muscle contraction". They find that when the three species and their annotations are lumped together there are 72 genes annotated to "cardiac muscle contraction" in at least one of the three species. They find that when looking at orthologs across those species it is very spotty which species has such annotation and which does not. They figured all the species genes and orthologs should have these annotations. Attached is a spreadsheet of the data he shared with me. I'm wondering if there is a way to find out which of the associated gene families may have gone through PAINT and when. Perhaps at least some of these haven't been "PAINTed"? I'm not sure if they considered only direct annotation to "cardiac muscle contraction" or if they considered children as well...
Link to the spreadsheet with the comparison:
https://docs.google.com/spreadsheets/d/1mHtkdto4qD9LIMHoaE_N016EBFlPOX3lFJUVLAtUmHQ/edit?gid=0#gid=0

[thomaspd]

I've looked at the list, and here are my observations:
It appears that the user only looked at annotations directly to this term (not child terms), and that the PAN-GO (evidence code IBA) annotations are consistent across zebrafish, mouse and human orthologs. The apparent inconsistencies are due to a couple of different factors:

1. Many apparent inconsistencies are just due to 1:many orthologs. Gene duplication in the teleosts after divergence from the tetrapods results in cases where there are multiple paralogs in zebrafish that are all orthologs of a single mouse or human gene. An example is tnn2a.1, tnn2a.2, tnn2a.3, tnn2a.4, tnn2b.1 and tnn2b.2 in zebrafish—these all have a single ortholog in mouse and human (TNN2).
2. Differences in experimental (not PAN-GO) annotations for the mouse, human and zebrafish orthologs. This is common for experimental (evidence code EXP, IDA, IMP, etc.) GO annotations, as it reflects the differences in published experimental results for orthologs in different species/experimental systems.

[cmungall]

I think the numbers will make more sense after the next release in a few days; see:

• Changes in PAINT data - July 2024 go-releases#92