17 May 2023 PAINT Conference Call

From GO Wiki
Jump to navigation Jump to search

Present

Marc, Dustin, Anushya, Paul T, Huaiyu, Pascale


Agenda

Dropping mouse and rat IMP annotations from the PAINT load (by Pascale)

  • Marc found that families with genes involved in immune response have a large number of IMP annotations that are usually not accurate. These annotations affect the efficiency in paint curation.
  • The decision is to still load all experimental annotations to the PAINT data. The PAINT tool will be modified to display an evidence code filter page so that the curator can select experimental annotations with specific code to display. The matrix will only show the experimental annotation with selected ECs. Annotations with unselected ECs will not be included in the matrix.

PANTHER 18 build

Node forward tracking comparisons

  • 15.0 -> 18.0 (Note: these 15->18 stats are incorrect! Refer to the 14_June_2023_PAINT_Conference_Call#PANTHER_18_build for the correct stats)
    • LEAF - 1776325 / 1968819 = 90.2%
    • SPECIATION - 1073649 / 1138527 = 94.3%
    • DUPLICATION - 291797 / 367319 = 79.4%
    • HORIZONTAL TRANSFER - 5185 / 6838 = 75.8%
    • ALL - 3146956
  • 15.0 -> 17.0
    • LEAF - 1847433 / 1973248 = 93.6%
    • SPECIATION - 1025670 / 1137262 = 90.1%
    • DUPLICATION - 299119 / 371052 = 80.6%
    • HORIZONTAL TRANSFER - 4436 / 6973 = 63.6%
    • ALL - 3176658
  • 17.0 -> 18.0
    • LEAF - 1887652 / 1968819 = 95.8%
    • SPECIATION - 1078875 / 1138527 = 94.7%
    • DUPLICATION - 309757 / 367319 = 84.3%
    • HORIZONTAL TRANSFER - 5374 / 6838 = 78.6%
    • ALL - 3281658

Based on the review of some trees, it seems that the v18 trees are more similar to v15 trees, which is consistent with the results above.

  • It seems that the trees map better from 15 to 18, pending results from 17 to 18 mapping. If that is the case, it makes more sense that we will migrate PAINT annotations from 15 to 17.
  • We will pause annotation review and curation in v17.
  • We will continue to gather results, especially the v17 to 18 map. We will also review more trees. A decision will be made based the results, probably in two weeks.

PAN-GO paper

Status update. What remains to be do?