Drug database for cell line experiments with before vs after treatment responses

Question

[fr]

Dear all,

I am looking for databases with cell line treatments with as many drugs as possible, and information from before vs after treatment, in terms of gene expression or similar. From what I see, CCLE has only treated data, CMap and LINCS/L1000 have before vs after data, but I was wondering if anyone is aware of something like this?

Thanks in advance

Answer 1

[svlachavas]

Hi rf,

your description in the question is a bit confused-if you actually want databases for drug repurposing or similar approaches, you have already mentioned LINCS/L1000-which has the L1000CDS2- (http://amp.pharm.mssm.edu/L1000CDS2/#/index)-from which you can query a subset of the L1000 data, as also some further options from Enrichr tool.

Best,

Efstathios

[fr]

I am looking for comprehensive experimental datasets for CLs exposed to drugs, with before vs after gene expression profiles. Specifically, I am looking for raw data, not connectivity measures or signature querying systems. I am aware of the CMap 02 and LINCS/L1000, which provide the raw data, but I am looking for additional databases

[svlachavas]

Again your answer is a bit misleading- about your concept of "raw data" - you mean you don't trust the signatures produced, and you want to do your own DE analysis ? this is for asking about raw data ? Or you want the actional gene-set libraries to perform some methodologies of your own ? The latter, is included in the libraries section for instance in Enrichr.

For additional tools/databases, check the following:

https://db.idrblab.org/ttd/

(For the actual raw data, i assume that you already know the most data are deposited as GEO datasets, regarding LINCS, etc)

[fr]

thanks for your input. Yes, I would like to get raw data (for instance .CEL files) so that I could process them myself. I do know about the GEO datasets, but unless I go through them one by one, it is hard to find largescale drug-response genesets that have before vs after treatment data (for instance, a few cell lines tested against a batch of drugs similarly to what CMap or LINCS did)

[svlachavas]

Rather challenging task in my opinion. Of course without a strong server or a cloud or something like this, to download all the raw data (from here i think more than https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE70138)- (http://www.lincscloud.org/)--and more importantly, you would need at least a beast machine to perform all kinds of downstream analysis-with the only exception if you could focus on one specific batch of pertubagens, etc.

In my opinion, it would be far more interesting the approach of ranking and scoring results from these gene-set libraries-as from my personal experience, the disease signatures they have produced, are valid.