I am looking to do an analysis of the communication network between two cell types. As such I am trying to create a list of secreted transcripts/genes and receptors.
Eliminating genes with null read counts is a first step. As a second step, I thought to take a look at signal recognition particles (SRP). Specifically, if the transcript for an SRP is not accounted for, it may be possible to discount the secreted genes which it chaperones to the ER.
I couldn't really find a good SRP-interaction database. Does one exist? Is this idea feasible or just completely wrong?
What other approaches could one take to eliminate/reduce the number of genes (aside from simply filtering out by function).
Note: (I realize that transcripts/genes they are in general different, but they are the same in the context of our experiment).
Thanks in advance.