iLINCS relation to CMap at clue.io

[Kinnaird, Brian T]

Hello,

I'm a medical student who is looking at the signatures for antidepressant perturbagens in the NPC and NEU cell lines. Looking at the perturbagens' signature IDs (e.g., "CPC015_NPC_24H:BRD-K70301876-034-06-1:10"), I notice there some of the signatures from the Connectivity Map at clue.io can also be found in iLINCS, but less than half. What is the relation between the data included in both tools? I was wondering if it is meaningful that a signature from the clue.io CMap is not represented in iLINCS. 

Also, you know what the difference is between signatures' IDs that have the prefix "CPC" versus "NMH"? I came across a few perturbagens that had two signatures with identical conditions except for the signature IDs were different; one started with CPC and the other was NMH001 or NMH002.

Brian Kinnaird, MA (he/him/his)

MD student, Class of 2023

University of Iowa Carver College of Medicine

mobile: (319) 329-7078 | email: brian-k...@uiowa.edu

[mario.me...@gmail.com]

Hello,

iLINCS contains original LINCS project L1000 signatures released through GEO and does not contain signatures that are specific to the newest Expanded LINCS library. We also apply a mild QC filtering on chemical perturbagen signatures and this could be another reason. Only chemical perturbagen signatures showing evidence of being reproducible by having the 75th quantile of pairwise Spearman’s correlations of level 4 replicates (Broad institute distil_cc_q75 quality control metric13) greater than 0.2 are included in the iLINCS library.

Please note that we are in the process of adding the rest of the signatures from the Expanded LINCS library and I post a message when this done.

Best,

Mario