pporrasebi
commented
4 years ago I cannot see the need for these terms. We record the expansion method as a separate column in the MITAB, the only format in which the expansion is really needed. What is the rationale behind needing to capture expansion methods in the interaction type field of our schemas? Is it for cases where there is a need to represent already expanded data and the non-expanded information is not available? If so, could you please give an example of a resonable use case, please?
These seem to be needed to formally annotate binary interactions resulting from spoke and matrix expanded complexes. It looks like they've got to go to either ' experimental interaction detection' (preferred) or their own branch as they cannot be made a part of 'inference' branch. This is b/c 'inference' is defined:
both expansion methods, when properly applied, result in the same binary interactions as these initially identified as binaries. if in doubt, squint your eyes and pretend for a while that only one, of the originally many, anti-prey blot was performed. the remaining binary interaction would be annotated as experimentally identified physical association. additional identification of another protein makes no difference - the binary interaction should remain annotated as experimentally identified physical association. the presence/absence of additional proteins(molecules) does make a difference on interaction type but not on the experimental method used to observe interaction
the point is that complex expansions are not based on human assumptions any more than experimentally identified binaries, experimental support is even more complete the for the binaries and there's no extension through homology. all of these disqualify complex expansions as inference methods, at least until the definition of inference is somehow modified.
if making complex expansions into subtypes of experimental methods is somehow controversial they can be made into their own 'complex expansion' sub-branch of 'interaction detection methods' defined as: 'complex expansion' - 'inference of binary interactions from experiment that originally detected association of more than two molecules
'spoke expansion' - 'inference of binary interactions between the bait and each prey molecule of affinity-detected multi-molecule association'
'matrix expansion' - inference of binary, physical interactions between each two molecules of experimentally-detected multi-molecule physical complex '
apart from these, yet another term can be created to annotate direct contacts between molecules in molecular complexes with experimentally-determined structures:
'structure expansion' (or 'topology expansion') - 'inference of binary, direct interactions between molecules found in direct contact in experimentally-determined structure of a multi-molecular complex'
the entire tree can be made either a child of 'interaction detection' or 'experimental interaction detection' term.