How do I determine which cell line(s) are most similar to a given cell line based upon the dependancies? This would be identical to the clustering of the genes, but on the other axis, so to speak.
Thanks for contacting DepMap. Unfortunately this is an intricate research question which goes beyond the scope of what we are able to respond to on the DepMap forum.
@DavidT We don’t have the results of such analysis available on the portal (and my guess is that it’d involve some non-trivial details to perform).
What do you think is the most exciting use case for it?
We have observed a few chemical genetic interactions in one cell line that are very robust, but aren’t seen in another related line. We would like to try to recapitulate this in another cell line. This may provide some way in which to guess which lines are most similar to ours and therefore most likely to mirror our initial observation.
Thanks Dave for sharing your application. You might want to check out Celligner on our portal as well. It uses gene expression rather than dependencies to represent samples and its focus is tumor/cell line alignment, but one could also use it to look for the cell lines most similar to a given cell line.