16 June 2015 PAINT Conference call

From GO Wiki
Jump to: navigation, search


PTHR23100 family Presentation

Rama presented family PTHR23100. She had 2 issues, otherwise the family was straightforward:

- question about horizontal transfer seemingly incorrect: eukaryotic version of gene is in plants and then in cyanobacteria. That seems to be backward. What to do when a tree seems to be incorrect? => send an email to Huaiyu.

- may be related to the previous point: localization in eukaryotes is mitochondrial (based on yeast annotation). In plants (Viridiplantae) localization is in chloroplast requiring a STOP and a propagation of the chloroplastic localization. This latter has to be STOP-ped again for the cyanobacteriall node under Viridiplantae. This again might suggest the tree is incorrect.

Other issues during the meeting

- Janos: In some organisms domains or 2 proteins that are sequential in the same pathway may fuse, while in others they remain separate. Is there any special procedure for cases like these? Paul: no special case, each sequence is in only one family. Where these sequences will show up is in families with more common function, like in kinases, ...

- which version of the software to use? Suzi: use the latest, 1.14, that is using v10 of Panther. The problem with NOT is not solved in the latest version. (In the annotation matrix (?) it appears as if lower level (more specific) terms are also STOP-ped when a high level term is STOP-ped. This is not true, check the annotations tab.)

- the issue from a recent email from Janos came up for discussion: inheritance of an annotation from a high level node cannot be blocked at an intermediate node if there is a direct annotation in a leaf node. A few arguments came up pro and con:

  * the annotation could be wrong, it should be fixed 
  * there is no way for a curator to indicate this part of the tree is not done, it should be revisited

=> Suzi committed to disable this check by end of June (or end of July the latest)

- Janos: can we use another symbol to make HGT more visible, the current diamonds are very hard to spot). Suzi: making them bigger would be an easy change to make.