Skip to main content
. 2021 Aug 20;7(34):eabh1275. doi: 10.1126/sciadv.abh1275

Fig. 6. Validation of clinical significance of predicted tumor dependencies by clinical data (TCGA) and preclinical drug trials (PDX Encyclopedia).

Fig. 6

(A) Predicted ESR1 dependencies in breast tumors (BRCA) with positive and negative immunohistochemical (IHC) status of estrogen receptor (ER+ and ER−). Statistical significance is assessed by the one-tailed t test. (B) ERBB2 dependencies in BRCA with different responses to trastuzumab. SD, stable disease. (C) Validation of predicted FGFR2/FGFR4 dependencies by drug response of LLM871 in preclinical PDX models. Predicted dependency scores of FGFR2 and FGFR4 were averaged for each PDX. Statistical significance is assessed by the Kruskal-Wallis test. (D) Predicted WRN dependencies across five MSI-prone cancers of TCGA, including colon adenocarcinoma, esophageal carcinoma, rectum adenocarcinoma, stomach adenocarcinoma, and uterine corpus endometrioid carcinoma. MSI categories were experimentally determined by TCGA: MSI-high (MSI-H), MSI-low (MSI-L), and microsatellite stable (MSS). Statistical significance is assessed by the one-tailed t test. See fig. S15 for the results of individual cancer types and pan-cancer MSI scores predicted by the MANTIS algorithm.