NTNU SKOR2 dispute Q2VWA4 GO:0000981 ECO:0000255 (ISM) PMID:19274049

RLovering commented 4 years ago

Hi Marcio (NTNU)

Q2VWA4 | GO:0000981 | ECO:0000255 (ISM) | PMID:19274049

Discussion so far

Ruth Lovering	07-MAY-2020	I don't think there is sufficient evidence to suggest this protein binds DNA. InterPro does not associated the GO term DNA binding with the InterPro record IPR009061 nor with IPR010919 IPR010919 record discusses how SKI (related to this gene) disrupts Tx:Ski is able to disrupt the formation of a functional complex between the Co- and R-SMADs, leading to the repression of TGF-beta, Activin and BMP responses, resulting in the repression of TGF-signalling Please remove the dbTF annotation Thanks Ruth

Marcio Luis Acencio | 11-MAY-2020 | Lambert et 2018 suggest that this protein is a DbTF due to direct evidence for its DNA binding in a HT-SELEX paper (PMID: 28473536). Maybe we can remove or modify this annotation to accommodate this experimental evidence. Ruth Lovering | 19-MAY-2020 | Hi Marcio, I have looked at this paper and checked the supplemental files https://science.sciencemag.org/content/sci/suppl/2017/05/03/356.6337.eaaj2239.DC1/aaj2239_Yin_SM.pdf you can search in the file for SKI etc but it is not listed, at least none of the synonyms for SKOR1 and SKI are in this list. To demonstrate try looking for ONECUT and if you increase the text size you will see ONECUT in that circular graph. Consequently I do not believe this paper provides any evidence that this family binds DNA. Please remove the annotation Thanks Ruth Marcio Luis Acencio | 19-MAY-2020 | Hi! This protein is SKOR2 and it is possible to find its PMW in the Table S2 (https://science.sciencemag.org/highwire/filestream/693922/field_highwire_adjunct_files/1/aaj2239_Yin_SM_tables_S1-S6.xlsx) of the paper. So I believe that we should either keep this annotation or change it to IDA supported by the Science paper.

RLovering commented 4 years ago

@colinlog

colinlog commented 4 years ago

It looks quite ok to me at first sight, although TAATTAA may also represent a favoured minor groove binding mode without much specificity. If Arttu Jolma sees a problem with this motif for SKOR2, he will tell me. Meanwhile, yes, annotation of the 544 human proteins for which this high throughput SELEX table of the Yin Science Paper provides a DNA-binding domain - target DNA motif relation in its S2 data sheet is something good to do. Even though there are no unique protein identifiers, there are geneIDs in the S1 file... ;-) I note that the Supplementary file does not contain the ARID factors that have the AT-hooks, and nor TBP, so quite clean with respect to GTFs and coregulators.

RLovering commented 4 years ago

Arttu Jolma confirms: SKOR2 seems to have a semi-decent SAND domain in it and although the binding site looks quite different from the other SANDs I would still give it a pretty decent chance for being a real TF. Yin et al paper also used E. Coli derived protein, so it can’t be a carryover of some specific TF from the lysates as with the Jolma 2013 things where we were using mammalian cell lysates.

So we will keep the dbTF annotations

Thanks for your patience

Ruth

RLovering commented 4 years ago

FYI PAINT has created all of these GO:0000981 | DNA-binding transcription factor activity, RNA polymerase II-specific based on expt from one member of the family (these are not yet loaded into P2G)

Thanks

Ruth

mlacencio commented 4 years ago

Hi @RLovering!

I have deleted ISM-based annotations for SKOR1 and SKOR2. I will create soon new annotations based on HT-SELEX papers.

Thanks

Best,

Marcio

RLovering commented 4 years ago

Hi Marcio

Thank you for confirming. Although I have already added SKOR1 and 2 DNA binding annotations based on that paper. I was also wondering if we should annotate all proteins listed in that paper to DNA binding. Colin has already identified all the UniProt IDs so these don't need to be found again.

Colin and Pascale are you happy to have all proteins identified by this HT-SELEX paper annotated to sequence-specific double-stranded DNA binding. Due to the way these experiments are performed and the necessity to clone each gene and express them individually in e.coli I think this should be IDA evidence.

Colin has already identified all the UniProt IDs in the supplemental table so this would be straightforward.

Best

Ruth

colinlog commented 4 years ago

Yes Ruth, the SELEX experiment can certainly demonstrate sequence specific-DNA binding (GO:0043565), and it is a direct assay, albeit high-throughput and performed at the hand of a robot, by all means.

pgaudet commented 4 years ago

All annotations from this paper have been removed. Thanks @mlacencio

geneontology / go-annotation

NTNU SKOR2 dispute Q2VWA4 GO:0000981 ECO:0000255 (ISM) PMID:19274049 #2983