Harmonizing across datasets to improve the transferability of drug combination prediction.

Combination treatment has multiple advantages over traditional monotherapy in clinics, thus becoming a target of interest for many high-throughput screening (HTS) studies, which enables the development of machine learning models predicting the response of new drug combinations. However, most existing models have been tested only within a single study, and these models cannot generalize across different datasets due to significantly variable experimental settings. Here, we thoroughly assessed the transferability issue of single-study-derived models on new datasets. More importantly, we propose a method to overcome the experimental variability by harmonizing dose-response curves of different studies. Our method improves the prediction performance of machine learning models by 184% and 1367% compared to the baseline models in intra-study and inter-study predictions, respectively, and shows consistent improvement in multiple cross-validation settings. Our study addresses the crucial question of the transferability in drug combination predictions, which is fundamental for such models to be extrapolated to new drug combination discovery and clinical applications that are de facto different datasets.

An indole diketopiperazine alkaloid and a bisabolane sesquiterpenoid with unprecedented skeletons from Aspergillus fumigatus.

Fumitryprostatin A (1), the first example of an indole diketopiperazine alkaloid with a tricyclic 5/6/5 skeleton characterized by a dipyrrolo[1,2-a:1',2'-d]pyrazine-5,10-dione ring system decorated with a prenylated indole moiety, and fuminoid A (2), a sesquiterpenoid with a bicyclo[3.2.1]octane ring featuring a novel carbon skeleton via the transformation of the methyl, were isolated from the fungus Aspergillus fumigatus along with six known diketopiperazine alkaloids. The structure with the absolute configuration of 1 was determined based on spectroscopic analyses and X-ray crystallographic analysis, while the configuration of 2 was assigned tentatively by 13C NMR data with DP4+ probability analyses and ECD calculations. A plausible biosynthetic pathway for 1 was proposed starting from L-Trp and L-Pro via normal indole diketopiperazine. Compound 1 exhibited moderate cytotoxic activity with an IC50 value of 14.6 µM, while compound 8 exhibited moderate immunosuppressive activity in vitro.