I'm struggling to figure out how to go from gene expression data to be able to create a network out of the most expressed genes in a cell line. Ok, I have developed an application for Cytoscape where I want to compare my analysis with research on the NCI-60 cell lines. My goal is to do these steps:

1: Use one of the cell lines in the NCI-60 cell lines, in my case I will use the NCI_H23 where there is 3 samples in this dataset.

2: Use the most expressed genes in the NCI_H23 cell line to create a subnetwork using a protein interaction network database like iRefIndex, etc.

After these two steps I should finally be able to upload the network to Cytoscape and perform my analysis, but being a programmer I'm kind of stuck. I have tried using one of the 3 samples where I mapped the Affymatrix IDs to gene symbol and then selecting the top 20% genes with the highest values using a cutoff value, but I'm not sure if this is the correct way to do it and I'm not sure how to do step nr.2 above if my list of genes is correct.

My application investigates drug combinations based on graph theory so I want to create a network that represents a cell-line so I can compare my scoring of combinations with the ones from the NCI Almanac study. I don't want to compare it with other cell-lines, but rather find the most expressed genes in that cell-line to then retrieve interaction data from iRefIndex to get edges for the network. I guess I should average the 3 samples first before I repeat what I listed above.

I would really appreciate help on this as I've been stuck on this for over a week now.

What I'm stuck at now is trying to create a network from iRefIndex (using iRefR) using the set of gene symbols that I have extracted from the 3 samples. To be honest, I'm finding it hard to find the right tools and assessing what kind of interaction data i should use from iRefIndex etc, so I guess I am looking for confirmation that what I'm trying makes sense.

  • $\begingroup$ You seem to make quite a lot of guesses, maybe you should seek guidance with an advisor (offline). Usually one selects the genes that have a change in the expression when comparing with and without the drug. Have you read how other people investigate drug combinations? I don't know why have you selected the iRefIndex database, but are you looking for a interaction database or a pathway database? $\endgroup$
    – llrs
    Jan 18 '18 at 7:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.