search

MaayanLab / sigcom-lincs

Signature Commons LINCS Repo
3 stars 3 forks source link

LINCS L1000 Chemical Perturbations #68

Open wangjiawen2013 opened 1 year ago

wangjiawen2013 commented 1 year ago

Hi, Thanks for your work. I queried some up and down-regulated genes on sigcom web, there were some results from different data source, and here are two of them: image

Could you tell what's the difference between "LINCS L1000 Chemical Perturbations (2021)" and "LINCS L1000 Chemical Perturbations Consensus Signatures" highlighted by the red rectangle? Besides, the search results table of the first one includes details such as dose, tissue, cell line, time points, while the latter lacks of these details, where to find them ?

AviMaayan commented 1 year ago

@wangjiawen2013 thank you for your question... the "LINCS L1000 Chemical Perturbations Consensus Signatures" is our attempt to combine the signatures across all conditions into a consensus up and down genes for each drug. So it is the same data, but in the latter, you have one entry for each drug. This is useful for various use cases.

wangjiawen2013 commented 1 year ago

Thanks, This is also what I guessed. I found consensus signatures worked better for my query. Does "the signatures across all conditions" include different cell lines and drug doses and time point ?

AviMaayan commented 1 year ago

Yes.

wangjiawen2013 commented 1 year ago

Hi, sorry to trouble you again. Though it worked better using consensus sigatures, I still don't know the reason. Why the consensus signature performed better ? what is the use cases it suitable for ?

AviMaayan commented 1 year ago

Hi @wangjiawen2013 it likely worked better because your input is from cells that are not profiled with the L1000 data. So the consensus captures better the cell agnostic effects of the preturbagen.