AmiGO stats count differently than geneontology.org and release.go statistics

[suzialeksander]

AmiGO seems to be counting differently than the numbers shown on release/current and the GO homepage stats. Looking at EXP, it seems that AmiGO includes High Throughput codes with a count of 998,484:

http://amigo.geneontology.org/amigo/search/annotation?q=*:*&fq=evidence_subset_closure_label:%22experimental%20evidence%22&sfq=document_category:%22annotation%22

But http://geneontology.org/stats.html and the release list “Annotations for evidence EXP" 939,436

Just noting this discrepancy for now, not sure if stats should be changed or this just needs to be documented (but where?).

Probably what @pgaudet noted at the bottom of #598

[suzialeksander]

@dustine32 found the code:

https://github.com/geneontology/go-stats/blob/89ef0e2238b16359361ed47cd938c24d526e4971/libraries/go-stats/go_stats_utils.py#L36-L43

EVIDENCE_GROUPS = {
"EXP": ["EXP", "IDA", "IEP", "IGI", "IMP", "IPI"],
"HTP": ["HDA", "HEP", "HGI", "HMP", "HTP"],
"PHYLO": ["IBA", "IRD", "IKR", "IMR"],
"IEA": ["IEA"],
"ND": ["ND"],
"OTHER": ["IC", "IGC", "ISA", "ISM", "ISO", "ISS", "NAS", "RCA", "TAS"]
}