Responder.bio

Attribution

Data sources.

Everything on this site is computed from public cell-line pharmacogenomic data. Those datasets are licensed CC BY 4.0, which requires attribution wherever the data or anything derived from it is published. This is that attribution.

Required attribution

Cell-line viability data from the Cancer Dependency Map (DepMap), Broad Institute. Compound viability data: PRISM Repurposing 19Q4, doi:10.6084/m9.figshare.9393293. Expression and CRISPR (Chronos) gene effect data: DepMap 23Q2 Public, doi:10.6084/m9.figshare.22765112. Both licensed CC BY 4.0 (creativecommons.org/licenses/by/4.0/). Data were modified: subset, reshaped and aggregated. This work is not endorsed by the Broad Institute.

The releases, and how we know which they are.

Both releases were identified against the publishers' own file manifests rather than assumed. The DepMap 23Q2 Public files match the published record on size and MD5 exactly. The PRISM Repurposing 19Q4 identification is evidence-backed rather than checksum-proven, because the file we work from is a reshaped subset rather than the file as downloaded — so no published checksum could match it.

DepMap 23Q2 Public

doi:10.6084/m9.figshare.22765112

Expression and CRISPR (Chronos) gene effect

Verified by exact MD5 match

CC BY 4.0

PRISM Repurposing 19Q4

doi:10.6084/m9.figshare.9393293

Compound viability (primary screen)

Evidence-backed identification

CC BY 4.0

What “modified” means here.

We do not redistribute these datasets. Every number on this site is a derivative: the response matrix is subset to a fixed cell-line panel and reshaped, responses are aggregated over cell-line subgroups, and predictions for compounds absent from the measured data are computed from chemical structure. CC BY 4.0 permits this and requires that we say so.

What this is not.

Cell-line data is a triage substrate, not a clinical one. Everything here is hypothesis-generating and must be validated experimentally before it is relied on. Nothing on this site is a clinical or efficacy prediction, and none of it is medical advice.