Causal Graph

Try: KRAS G12D colorectal BRCA1 breast cancer EGFR NSCLC BRAF V600E melanoma

Loading causal graph…

Node Types

Gene
Pathway
Drug candidate
SAE feature ? Sparse Autoencoder features — mechanistic interpretability signals identifying which internal AI model features activate for each gene-drug relationship.

Experimental. Not populated for most analyses. (hidden)

Edge Types

Activates
Suppresses
Modulates
⊡
+
−
⊙
Select a gene, pathway, or drug node to see details, evidence, and connections.