From Machine Learning to Natural Product Derivatives that Selectively Activate Transcription Factor PPARγ

Abstract

Advanced kernel‐based machine learning methods enable the identification of innovative bioactive compounds with minimal experimental effort. Comparative virtual screening revealed that nonlinear models of the underlying structure–activity relationship are necessary for successful compound picking. In a proof‐of‐concept study a novel truxillic acid derivative was found to selectively activate transcription factor PPARγ.

Pre-Processing Structured Data for Standard Machine Learning Algorithms by Supervised Graph Propositionalization - A Case Study with Medicinal Chemistry Datasets

“…It has been suggested in the past, e.g., in [3,25], that using the domain specific feature sets in conjunction with the machine learning approaches could help increasing the classifier accuracy. Therefore in addition to the features generated by the propositionalization methods, we also have used publicly available chemical descriptors as input features in conjunction with the generated features.…”

Discovery of Natural Products that Modulate the Activity of PPARgamma: A Source for New Antidiabetics

“…Eight compounds exhibit agonistic activity towards PPARα, PPARγ or both. The most potent PPARγ-selective hit was a derivative of the natural product truxillic acid [97]. Using a pharmacophore-based VS of 19,892 natural products, Fakhrudin and coworkers [98], identified several neolignans, such as dieugenol, tetrahydrodieugenol and magnolol, as PPARγ partial agonists.…”

“…Of the eight compounds tested, two were derived from the natural compound α-santonin and were able to promote the PPARγ transactivation activity in a cell-based reported ADME absorption, distribution, metabolism and excretion; PPAR proliferator-activated receptor; VS virtual screening Table 6.3 Several successful examples of VS procedures used for identifying PPARγ agonists among natural products or derivatives gene assay, with values of 31 and 8 % for maximal PPARγ activation relative to pioglitazone [96]. Rupp and coworkers [97] combined several machinelearning methods to virtually screen a database of 360,000 compounds. They tested 15 compounds in a cellular reporter assay [97].…”

“…Rupp and coworkers [97] combined several machinelearning methods to virtually screen a database of 360,000 compounds. They tested 15 compounds in a cellular reporter assay [97]. Eight compounds exhibit agonistic activity towards PPARα, PPARγ or both.…”

Target Profile Prediction: Cross‐Activation of Peroxisome Proliferator‐Activated Receptor (PPAR) and Farnesoid X Receptor (FXR)

“…[13] We tested our target profile prediction software LeadHopper [14] for its ability to recognize known targets of GW4064 and correctly predict whether PPAR agonists 1 and 2 have the potential to activate FXR. In a previous study, this prediction tool correctly identified hitherto unknown targets of modulators of metabotropic glutamate receptors.…”