Research Use Only. KIRhub outputs are computational research artifacts. They are not validated for clinical decision-making, diagnosis, or treatment.

← The Pathway Atlas

The pharma-tumor co-embedding

Derived

14,118 tumors AND 92 drugs in one map. Tumors plotted by G[t]; drugs plotted by −Π_d so that spatial proximity literally means "this drug's perturbation reverses this tumor's pathway state."

PC1 (0.0% var)PC2 (0.0% var)AbemaciclibAbrocitinibAcalabrutinibAfatinibAlectinibAlpelisibApatinibAsciminibAvapritinibAxitinibBaricitinibBinimetinibBosutinibBrigatinibCabozantinibCanertinibCapivasertibCapmatinibCeritinibCobimetinibCrizotinibDabrafenibDacomitinibDarovasertibDasatinibDefactinibDeucravacitinibDuvelisibEncorafenibEntrectinibErdafitinibErlotinibEverolimusFedratinibFostamatinibFutibatinibGedatolisibGefitinibGilteritinibIbrutinibIdelalisibImatinibInavolisibInfigratinibLapatinibLarotrectinibLazertinibLeniolisibLenvatinibLorlatinibMidostaurinMitapivatMobocertinibNeratinibNetarsudilNilotinibNintedanibOsimertinibPacritinibPalbociclibPaxalisibPazopanibPemigatinibPexidartinibPirtobrutinibPonatinibPralsetinibQuizartinibRabusertibRegorafenibRemibrutinibRepotrectinibRibociclibRipretinibRuxolitinibSelpercatinibSelumetinibSirolimusSorafenibSunitinibTemsirolimusTenalisibTepotinibTivozanibTofacitinibTrametinibTucatinibUmbralisibUpadacitinibVandetanibVemurafenibZanubrutinib

Hide tumors via the legend to see drug-class structure alone. Hide drug classes to see where the unmet biology lives.

Next steps

Quantify the deserts per cohort → · Find drugs whose ideal patient doesn't exist →