rachhuntley
commented
6 years ago Putting this here for reference. I wanted to check with experts that the proposal here for the two child terms was acceptable; this is the communication I had with Sylvie Ricard-Blum from MatrixDB (member of IMEx Consortium)
Dear Rachel, Thank you very much for your e-mail. I am working on the extracellular matrix for more than 30 years and my team builds and analyzes extracellular matrix interaction networks for a long time. We use GO terms on a regular basis to contextualize these networks and I agree with you that the GO terms under "extracellular matrix" are not ideally representative at the moment.
It is an excellent idea to use the Hynes lab’s Matrisome Project, which is very well done. I know this work, and Richard Hynes and Alexandra Naba (the young scientist involved in the project) and we use their classification to analyze the networks we build.
I agree on both terms you suggested "proteinaceous extracellular matrix core" and "proteinaceous extracellular matrix core-associated".
Further changes should be done in my opinion to optimize GO terms under "extracellular matrix" and terms annotating extracellular matrix proteins or proteoglycans.It is a concern because we use them to build specific networks from the interaction database focused on the extracellular matrix we have developed (MatrixDB http://matrixdb.univ-lyon1.fr/). I would thus be delighed to participate in this reorganization if my expertise is helpful. I will look to the Github project of course but would appreciate to discuss with you at some point if you are interested in moving further on the extracellular matrix.
Best wishes. Sylvie.
Professor Sylvie RICARD-BLUM Pericellular and Extracellular Supramolecular Assemblies UMR 5246 CNRS - University Lyon 1 Institute of Molecular and Supramolecular Chemistry and Biochemistry University Claude Bernard Lyon 1, 43 Boulevard du 11 novembre 1918 69622 Villeurbanne Cedex, France Tel : +33(0)4 72 44 82 32 E-mail: sylvie.ricard-blum@univ-lyon1.fr http://www.icbms.fr The French Society for Matrix Biology http://www.sfbmec.fr The International Society for Matrix Biology http://ismb.org
Le 05/03/2018 à 14:47, Huntley, Rachael a écrit : Dear Sylvie,
Birgit Meldal has suggested I contact your for your opinion on extracellular matrix protein classification?
I’m a Gene Ontology biocurator at University College London and we have been curating some ECM datasets, however I have recently noticed that the Gene Ontology structure under the ‘extracellular matrix’ term is poorly organised. I’m therefore trying to re-organise the GO terms under ‘extracellular matrix’ as they are not ideally representative at the moment.
I’d like to organise the metazoan extracellular matrix (called 'proteinaceous extracellular matrix' in GO) based on the work of the Hynes lab’s Matrisome Project (http://web.mit.edu/hyneslab/matrisome/). I’m not sure if you are aware of this work, but they use proteomics together with manual curation and computational methods to define categories of proteins as either “core matrisome” or “matrisome-associated”.
The core proteins they have defined are ECM glycoproteins, collagens and proteoglycans and the matrisome-associated proteins include ECM-remodeling factors, various secreted factors and other ECM-affiliated proteins, so this second list can be quite large (see attached spreadsheet for the proteins they have assigned to each category).
Based on this, we would have two GO terms: 'proteinaceous extracellular matrix core’ (to which only ECM glycoproteins, collagens and proteoglycans could be associated) 'proteinaceous extracellular matrix core-associated’ (to which the non-core ECM protein could be associated)
Does this sound like a reasonable approach to you? Do you see any problems with the way the Hynes lab have categorised the proteins?
I have started a GitHub project to request the ontology changes I would like to see. If you are interested you are welcome to add further suggestions or comments: https://github.com/geneontology/go-ontology/projects/9
Many thanks in advance for your help. Best wishes, Rachael.
I’d like to reflect that there are a core set of well-known types of proteins present in the metazoan ECM ('proteinaceous extracellular matrix' GO:0005578), which have been determined based on their hallmark domain-based structure, and then many other proteins that are associated with the core ECM. This is based on the work of the Hynes group (http://matrisomeproject.mit.edu/) (and backed by ECM expert Manuel Mayr).
I propose to distinguish between these using two GO terms: o NTR: proteinaceous extracellular matrix core (part_of proteinaceous extracellular matrix). Definition: The proteinaceous extracellular matrix core consists of a restricted set of proteins that assemble to form an extracellular matrix. This includes collagens, proteoglycans and extracellular matrix glycoproteins (PMID:21937732, 23199376, 22159717). Synonyms: core matrisome, core ECM, core extracellular matrix.
o NTR: proteinaceous extracellular matrix core-associated (part_of proteinaceous extracellular matrix). Definition: The part of the proteinaceous extracellular matrix excluding the core. It includes i) extracellular matrix-affiliated proteins (affiliated either structurally or physically with the core matrisome), ii) ECM-remodeling enzymes, such as matrix metallopeptidases and iii) secreted factors such as growth factors and cytokines, known or suspected to bind to ECM. (PMID: 22159717). Synonym: matrisome-associated, ECM-associated, extracellular matrix-associated.
See GitHub #15294