PT  - JOURNAL ARTICLE
AU  - Lisa-Katrin Schätzle
AU  - Ali Hadizadeh Esfahani
AU  - Andreas Schuppert
TI  - Methodological Challenges in Translational Drug Response Modeling in Cancer
AID  - 10.1101/731836
DP  - 2019 Jan 01
TA  - bioRxiv
PG  - 731836
4099  - http://biorxiv.org/content/early/2019/08/12/731836.short
4100  - http://biorxiv.org/content/early/2019/08/12/731836.full
AB  - Translational models directly relating drug response-specific processes observed in vitro to their in vivo role in cancer patients constitute a crucial part of the development of personalized medication. Unfortunately, ongoing research is often confined by the irreproducibility of the results in other contexts. While the inconsistency of pharmacological data has received great attention recently, the computational aspect of this crisis still deserves closer examination. Notably, studies often focus only on isolated model characteristics instead of examining the overall workflow and the interplay of individual model components. Here, we present a systematic investigation of translational models using the R-package FORESEE. Our findings confirm that with the current exploitation of the available data and the prevailing trend of optimizing methods to only one specific use case, modeling solutions will continue to suffer from non-transferability. Instead, the conduct of developing translational approaches urgently needs to change to retrieve clinical relevance in the future.