Skip to main content

Table 1 Results from the cross-validation

From: Dose–response prediction for in-vitro drug combination datasets: a probabilistic approach

   

Latent GP

Dose response

 

\(r_d\)

\(r_c\)

wRMSE

Pearson’s r

wRMSE

Pearson’s r

PIICM

200

10

0.3767

0.5336

0.0680 (0.0724)

0.9800 (0.9733)

PIICM (no invariance)

100

10

1.1231

0.1895

0.1857 (0.0724)

0.8815 (0.9733)

PIICM (no measurement error)

200

10

1.6459

0.1731

0.2120 (0.0724)

0.8349 (0.9733)

  1. Results are shown for prediction in the latent space, as well as for the resulting prediction of dose response. Corresponding results on dose–response for the baseline model are shown in parentheses for each setting. For both the latent GP and the dose–response, we compare our predictions against estimated values obtained from each experiment’s raw viability measurements using the bayesynergy R package. Bold numbers indicate best performance