Geschmack ist vorhersagbar: Mit FlavorMiner. FlavorMiner heißt das Tool, das IPB-Chemiker und Partner aus Kolumbien jüngst entwickelt haben. Das Programm kann, basierend auf maschinellem Lernen (KI), anhand der…
Seit Februar 2021 bietet Wolfgang Brandt, ehemaliger Leiter der Arbeitsgruppe Computerchemie am IPB, sein Citizen Science-Projekt zur Pilzbestimmung an. Dafür hat er in regelmäßigen Abständen öffentliche Vorträge zur Vielfalt…
A bottleneck in the development of new anti-cancer drugs is the recognition of their mode of action (MoA). We combined metabolomics and machine learning to predict MoAs of novel anti-proliferative drug candidates, focusing on human prostate cancer cells (PC-3). As proof of concept, we studied 38 drugs with known effects on 16 key processes of cancer metabolism, profiling low molecular weight intermediates of the central carbon and cellular energy metabolism (CCEM) by LC-MS/MS. These metabolic patterns unveiled distinct MoAs, enabling accurate MoA predictions for novel agents by machine learning. We validate the transferability of MoA predictions from PC-3 to two other cancer cell models and show that correct predictions are still possible, but at the expense of prediction quality. Furthermore, metabolic profiles of treated cells yield insights into intracellular processes, exemplified for drugs inducing different types of mitochondrial dysfunction. Specifically, we predict that pentacyclic triterpenes inhibit oxidative phosphorylation and affect phospholipid biosynthesis, as supported by respiration parameters, lipidomics, and molecular docking. Using biochemical insights from individual drug treatments, our approach offers new opportunities, including the optimization of combinatorial drug applications.
Publikation
Saoud, M.; Grau, J.; Rennert, R.; Mueller, T.; Yousefi, M.; Davari, M. D.; Hause, B.; Csuk, R.; Rashan, L.; Grosse, I.; Tissier, A.; Wessjohann, L. A.; Balcke, G. U.;Advancing anticancer drug discovery: leveraging metabolomics and machine learning for mode of action prediction by pattern recognitionAdvanced Science112404085(2024)DOI: 10.1002/advs.202404085
A bottleneck in the development of new anti‐cancer drugs is the recognition of their mode of action (MoA). Metabolomics combined with machine learning allowed to predict MoAs of novel anti‐proliferative drug candidates, focusing on human prostate cancer cells (PC‐3). As proof of concept, 38 drugs are studied with known effects on 16 key processes of cancer metabolism, profiling low molecular weight intermediates of the central carbon and cellular energy metabolism (CCEM) by LC‐MS/MS. These metabolic patterns unveiled distinct MoAs, enabling accurate MoA predictions for novel agents by machine learning. The transferability of MoA predictions based on PC‐3 cell treatments is validated with two other cancer cell models, i.e., breast cancer and Ewing\'s sarcoma, and show that correct MoA predictions for alternative cancer cells are possible, but still at some expense of prediction quality. Furthermore, metabolic profiles of treated cells yield insights into intracellular processes, exemplified for drugs inducing different types of mitochondrial dysfunction. Specifically, it is predicted that pentacyclic triterpenes inhibit oxidative phosphorylation and affect phospholipid biosynthesis, as confirmed by respiration parameters, lipidomics, and molecular docking. Using biochemical insights from individual drug treatments, this approach offers new opportunities, including the optimization of combinatorial drug applications.
Publikation
Gago-Zachert, S.; Schuck, J.; Weinholdt, C.; Knoblich, M.; Pantaleo, V.; Grosse, I.; Gursinsky, T.; Behrens, S.-E.;Highly efficacious antiviral protection of plants by small interfering RNAs identified in vitroNucleic Acids Res.479343-9357(2019)DOI: 10.1093/nar/gkz678
In response to a viral infection, the plant’s RNA silencing machinery processes viral RNAs into a huge number of small interfering RNAs (siRNAs). However, a very few of these siRNAs actually interfere with viral replication. A reliable approach to identify these immunologically effective siRNAs (esiRNAs) and to define the characteristics underlying their activity has not been available so far. Here, we develop a novel screening approach that enables a rapid functional identification of antiviral esiRNAs. Tests on the efficacy of such identified esiRNAs of a model virus achieved a virtual full protection of plants against a massive subsequent infection in transient applications. We find that the functionality of esiRNAs depends crucially on two properties: the binding affinity to Argonaute proteins and the ability to access the target RNA. The ability to rapidly identify functional esiRNAs could be of great benefit for all RNA silencing-based plant protection measures against viruses and other pathogens.
Publikation
Nettling, M.; Treutler, H.; Cerquides, J.; Grosse, I.;Unrealistic phylogenetic trees may improve phylogenetic footprintingBioinformatics331639-1646(2017)DOI: 10.1093/bioinformatics/btx033
MotivationThe computational investigation of DNA binding motifs from binding sites is one of the classic tasks in bioinformatics and a prerequisite for understanding gene regulation as a whole. Due to the development of sequencing technologies and the increasing number of available genomes, approaches based on phylogenetic footprinting become increasingly attractive. Phylogenetic footprinting requires phylogenetic trees with attached substitution probabilities for quantifying the evolution of binding sites, but these trees and substitution probabilities are typically not known and cannot be estimated easily.ResultsHere, we investigate the influence of phylogenetic trees with different substitution probabilities on the classification performance of phylogenetic footprinting using synthetic and real data. For synthetic data we find that the classification performance is highest when the substitution probability used for phylogenetic footprinting is similar to that used for data generation. For real data, however, we typically find that the classification performance of phylogenetic footprinting surprisingly increases with increasing substitution probabilities and is often highest for unrealistically high substitution probabilities close to one. This finding suggests that choosing realistic model assumptions might not always yield optimal predictions in general and that choosing unrealistically high substitution probabilities close to one might actually improve the classification performance of phylogenetic footprinting.