Which methods have been proved to be the most reliable in detecting PPIs?

Question

[GD]

I downloaded the current BIOGRID-ORGANISM-Saccharomyces_cerevisiae_S288c-3.4.132.mitab and I see that:

1. There are different methods to detect PPIs, such as: FRET, PCA, Co-fractionation, Co-purification [...]; and
2. There are entries with no confidence-score provided ("-" at column 15, see mitab-format here).

Given (2), my question is: Which methods have been proved to be the most reliable in detecting PPIs?

[istloui1]

How the BioGRID score is derived?

Answer 1

[Jean-Karim Heriche]

In my opinion, no method is overall better than any other, they just have different scopes and limitations. In my work, I don't find this kind of scoring useful but I guess it depends on what you're trying to do. If you're concerned about false positive interactions, the best way to remove them is to look for independent validation e.g. an interaction seen with more than one method or reported in more than one paper (from different authors). Note also that BioGRID scores originate from the papers describing the interactions so they are not directly comparable across studies (some would be p-values, others confidence scores of various sorts). Finally, individual databases are usually not exhaustive, you get better coverage by combining interactions from multiple databases like iRefIndex does.

[GD]

Very informative!

Thank you so much, :) .

[GD]

What about the Multi-validated (MV) datasets from BioGRID?

Still single database, but multi-validated.

[Jean-Karim Heriche]

That goes some way to fit the criteria for independent validation but they don't seem to track authors: same experimental system with different publication sources could be different papers by the same group reporting the same interaction using the same experimental set up. Checking for independent validation is not easy and likely results in rejecting many good interactions. As usual, the acceptable trade off between false positives and false negatives depends on the specifics of the problem at hand.