Proteochemometrics and drug development

The focus of our research is to aid the drug discovery process with computer-aided decision support using the proteochemometric method, and also to develop new drugs based on natural products.

Principal investigator: Professor Jarl Wikberg

Highlighted projects:

(I) Pharmacology of the libiguins

Studies on the libiguins ; in part a collaboration with Philippe Rasoanaivo, IMRA, Antananarivo, Madagascar, Aigars Jirgensons, IOS, Riga, Latvia, Richard Svensson, inst. F. Farmaci, Uppsala University.

Libiguins are limonoid terpenoid compounds originally discovered in a Malagasy mahogany species, Neobeguea mahafalensis, which show unprecedented powerful and potent stimulatory effects on sexual behavior (see Wikberg et al, Planta Med. 2014 Mar;80(4):306-14. doi: 10.1055/s-0033-1360390). The libiguins are available only in minute quantities naturally, but we have developed a highly efficient methods for their synthesis allowing them to be produced in large quantities (Grigorjeva et al, J Org Chem. 2014 May 2;79(9):4148-53. doi: 10.1021/jo500318w.). Using the synthetic libiguins the pharmacology and mechanism of action of the libiguins are now under study. The aim of the project is also to introduce the libiguins for treatment of sexual dysfunction and in particular erectile dysfunction in men.

(II) Isolation, structural determination and pharmacology of novel natural and semi-synthetic compounds

These studies are devoted to the isolation, structural determination and preliminary pharmacological characterizations of novel natural compound from various plants, primarily with origin from Madagascar. The project is a collaborations with  Philippe Rasoanaivo, IMRA, Antananarivo, Madagascar, and Torgils Fossen, Centre for Pharmacy, Department of Chemistry University of Bergen, Bergen, Norway.

(III) Proteochemometrics

Proteochemometric modeling is a bioactivity modeling technique to predict and analyze the interaction of drug candidates  (ligands, chemical compounds etc) with their target molecules (i.e. Macromolecules, proteins, receptors, enzymes, etc.). The technology is based on the description of both the drug candidates and the macromolecular targets with so called chemical descriptors and examples from measurements for drug-target interactions using machine learning in order to provide proteochemometric models from the combined data. Proteochemometric models can be used to predict the bioactivity of chemicals with target groups e.g. prior to the synthesis of the chemical compounds. This finds use in predicting novel drug candidates with improved properties and reduced side effects.


For a recent example for the use of proteochemometrics see e.g.  Lapins et al. PLoS One. 2013 Jun 17;8(6):e66566. doi: 10.1371/journal.pone.0066566.

The project is in part a collaboration with Chanin Nantasenamat, Center of Data Mining and Biomedical Informatics Faculty of Medical Technology, Mahidol University 999 Phutthamonthon 4 Road, Salaya, Nakhon Pathom