Recreate Table 2 showing descendant counts for depth-01 GO terms are highly skewed

[dvklopfenstein]

Currently, I wish to use the biological process groupings in the D1 Alias column of Table 2. However, I've had some trouble getting my counts to match.

My issues may be due to the (potentially depreciated?) goslim_generic file, but I cannot tell for sure. For instance, I get 12k terms in the cellular process group A instead of 19k shown in Table 2. Would you be able to share source code for Table 2? I would like to ensure my groupings are correct so that I can use this function

[dvklopfenstein]

I have created a script and added it to the GOATOOLS tests directory, tests/wr_tbl2_goatools_d1.py.

This script will exactly match the numbers in Table 2 if you use go-basic.obo that we used to generate the tables, located in the goatools_simulation repo. I removed many D1 GO terms so the table would fit for publication, but still transmit the idea that the depth-01 have highly skewed numbers of descendants.

It is likely that your difference of seeing 12k terms vs our 19k terms for cellular process is due to not using optional relationships, like part_of. We tried grouping with and without relationships and found that it worked much better and was far less laborious to create a sections file if we always use relationships for grouping.