PAINT - Apoptosis
This page captures PANTHER families with annotations to 'execution phase of apoptosis' GO:0097194. The definition is "A stage of the apoptotic process that starts with the controlled breakdown of the cell through the action of effector caspases or other effector molecules (e.g. cathepsins, calpains etc.). Key steps of the execution phase are rounding-up of the cell, retraction of pseudopodes, reduction of cellular volume (pyknosis), chromatin condensation, nuclear fragmentation (karyorrhexis), plasma membrane blebbing and fragmentation of the cell into apoptotic bodies. The execution phase ends when the cell has died." Thus, all signaling events *upstream from* the effector caspases (those include CASP3, CASP6, CASP7) should *not* be annotated to this term. Initiator caspases (e.g., CASP2, CASP8, CASP9, and CASP10) should also not be annotated to this term. See http://en.wikipedia.org/wiki/Caspase and http://en.wikipedia.org/wiki/Apoptosis for general background.
Apoptosis Project Report
In November 2013 we decided to assign families to annotators with the goal of setting clear deliverables. Having a common topic was expected to facilitate discussions among annotators. We chose apoptosis because the ontology and the annotations were recently revisited.
- The PAINT process includes the following steps (http://wiki.geneontology.org/index.php/PAINT_SOP):*
- Verification of the tree topology and composition
- Overview of literature on protein family function
- Ensuring sufficient annotation coverage
- Annotating ancestral genes
However until now most of the communication has been on the 'propagation' aspect of the annotation (step 4). As we go along, we realize that reviewing the primary annotations in light of a broad knowledge about a family is an essential step. This doesn't always result in changes in annotations, but if nothing else, it helps curators ensure that the primary annotations are correct and capture all the important aspect of a family's biological role. In some cases that review may uncover deeper problems with annotation of certain proteins, or with the interpretation of certain GO terms. We need to make sure that this 'integrative' aspect of the PAINT process gets well communicated both to members of the GOC and outside.
- Annotation review/quality control*
The most critical issue underlying incorrect annotations has to do with the understanding of the semantics of a GO annotation, ie, what does it mean when we associate a gene/protein with a GO term?
Specifically, those are common misinterpretations we encounter:
- Difficulty to distinguish between annotation to a process versus annotation to the regulation of that process. Because of the non-transitive nature of the 'regulation' relation, this has important implications for downstream data analysis.
- Difficulty to distinguish between experimental READOUT or ENDPOINT that is measured, and a real role in the process itself. This results in large number of over-annotations when proteins have an upstream/indirect role.
Out of 15 families, only 2 were correctly annotated to 'execution phase 'of apoptosis'. We need to evaluate the extent of this issue in the GO as a whole.
- The Protein2GO Dispute Mechanism* was very helpful as a way to discuss questionable annotations. We suggest that this get adopted as one of the tools for the PAINT workflow. Hopefully that will also help keep record of the disputed annotations and make stats with respect to which GO terms are more problematic.
- Ontology updates*
The Apoptosis PAINT curation project also resulted in several ontology updates; creation and obsoletion of terms; and addition of comments to help annotators use terms correctly.
Suggested action points from the apoptosis work:
- Change the emphasis of PAINT from 'phylogenetic annotation propagation' to 'Phylogenetic-based annotation integration'*
- Communicate those issues at an annotation call.
- Identify similar situations where annotators have a hard time distinguishing between readouts and processes and provide guidelines.
- Continue using PAINT as a mechanism for reviewing annotations.
'Please sign up for a family, and comment after you submit them to the SVN repository.'
|Family||PTHR ID||Curator||Execution phase of apoptosis annotation justified?|
|cell death activator cide||PTHR12306||Moni -- committed||No - negative regulation of execution phase of apoptosis|
|DNA fragmentation subunit Beta||PTHR13067||Moni -- committed||Yes - apoptotic DNA fragmentation|
|map-kinase activating death domain protein (madd)/denn/aex-3(c.elegans)||PTHR13008||Moni -- in progress|
|mitogen-activated kinase kinase kinase||PTHR24361||Huaiyu -- committed||No. PAKs are target of caspases, but not necessarily in the execution phase of apoptosis.|
|40s ribosomal protein s3a||PTHR11830||Pascale - committed||No - positive regulation of extrinsic apoptotic signaling pathway. Family also contains CYLD, a ubiquitin dehydrolase, that I annotated in Protein2GO on Dec 19 - still need to annotate "GO:0061578 Lys63-specific deubiquitinase activity" in PAINT|
|akt-related serine/threonine-protein kinase||PTHR24352||DHL||No.|
|bcl-2 related||PTHR11256||Huaiyu -- committed||No. Bcl2 related proteins are involved in the cytochrome c release from mitochondria, which leads to the activation of caspase cascade, so they are not in the execution phase of apoptosis.|
|bcl2/adenovirus e1b 19-kda protein-interacting protein||PTHR12112||Moni -- not committed.||Yes. BNIPL-2 interferes with GAP activity of Cdc42GAP toward Cdc42. Cdc42 recruits PAK1 and activates JNK1 to phosphorylate, which consequently inactivates the anti-apoptotic Bcl-2 protein, allowing p53-dependent apoptosis.|
|disulfide oxidoreductase||PTHR22912||Paul||No - caspase-independent apoptosis|
|ephrin||PTHR11304||Li -- committed||It's a "not" annotation to 'execution phase of apoptosis' GO:0097194. PMID:11870224 shows EphA3-expressing cell show morphological changes such as membrane blebbing. In contrast to apoptotic blebbs, which are static and often detach from the cell surface, ephrin-A5 triggers highly dynamic, spherical membrane protrusions that do not lead to cell fragmentation. "not" apoptosis process instead?|
|f-box and wd40 domain protein||PTHR22844||Rama|
|family not named||PTHR18834||Li -- committed||No - SRA overexpression enhanced apoptosis in mammary epithelia|
|family not named||PTHR24025||Karen - not well enough annotated to do PAINT||No - Dsg2 is a target of caspase 3, not playing a role in the execution phase itself|
|group ii pyridoxal-5-phosphate decarboxylase||PTHR11999||Rama|
|microtubule-associated protein||PTHR13843||DHL||Yes - microtubule cytoskeleton organization|
|protein arginine n-methyltransferase||PTHR11006||Pascale -- committed||No (the paper shows that "Transfection of PRMT2 into 293 cells increased their susceptibility to TNF-induced cell death"; is this really sufficient evidence for an annotation?)|
|protein disulfide isomerase||PTHR18929||Paul||No - regulation of apoptosis|
|rbm25 protein||PTHR18806||Pascale - committed||No - regulation of alternative mRNA splicing, via spliceosome. Note that it regulates BCL-x spilcing.|
|telomeric repeat binding protein||PTHR21717||Li -- committed||No, regulation term is better choice or just apoptosis process|
|tumor protein d52||PTHR19307||Karen - not well enough annotated to do PAINT||No - not execution phase, regulation of apoptotic signalling|
|type i interferon||PTHR11691||Pascale - commited||No, should be regulation of apoptosis (human IFNB1)|
|u3 small nucleolar rna-associated protein 11||PTHR12838||Karen - committed||No - experiment supports "positive regulation of apoptotic process" at best; however, fairly old non-specific experiment and this family is now known to be the Utp11 component of the small subunit processome, and that disruption of ribosomal biogenesis and nucleolar structure leads to activation of cell cycle arrest and/or apoptosis; not clear that role of Utp11 in apoptosis is specific|
|ubiquilin||PTHR10677||DHL||No, regulation of protein degradation.|
To do next: Families annotated to children of 'execution phase of apoptosis'
|Family||PTHR ID||Curator||Execution phase of aopoptosis annotation justified?|
|glucagon||PTHR11418||Pascale -- commited||No - negative regulation of apoptosis|
|30s/40s ribosomal protein s3||PTHR11760||Pascale -- commited||No - regulation of apoptosis|
|rho-type gtpase activating protein||PTHR12659||Moni -- in progress|
|family not named (BNIP3)||PTHR15186||Pascale -- committed||No - In fact, apoptosis is wrong here. BNIP3 mediate caspase-independent programmed cell death|
|cop9 signalosome complex subunit 7/dendritic cell protein ga17||PTHR15350||Pascale -- committed||No - Mutant has defect in apoptosis; regulation would be better|
|myosin light chain kinase-related||PTHR22964|
|paraneoplastic antigen||PTHR23095||Pascale - committed||No - regulation|
|translation factor||PTHR23115||Pascale - Will not do yet; translation factor, big family, all factors should be done at same time||No|
|family not named||PTHR24271||Li -- committed||yes|
|family not named||PTHR24272||Moni -- in progress|
|family not named (PTGIS)||PTHR24295 - now PTHR24306||Pascale -- committed||No, regulation of apoptosis|
|family not named||PTHR24377||Pascale -- committed||No, regulation of apoptosis // Zinc finger - huge family 2000 proteins|
|family not named||PTHR32047||Li -- committed||no, PMID:11466612 shows activities of caspase-3 like proteases enhanced twofold in Bcl10 transfected 293T cells, further experiments show Bcl10 may involved in apoptosis signaling by alter its binding partners...|