A repository of structures, experimental data and QSAR models for molecules with antitumor activity.
RSS icon Email icon Home icon
  • About the repository

    Welcome to the anticancer.irb.hr repository with datasets of compounds showing activity against various cancer cell lines in vitro. We used them for QSAR studies that predict antiproliferative activity from structural features of the molecules, as well as for determining the putative biological mechanism from the resistance or sensitivity patterns across cancer cell lines.

    Antitumor and antibacterial activity of crown ethers. Data available here >>

    • Supek F et al.Could LogP be a principal determinant of biological activity in various 18-crown-6 ethers?  Synthesis of biologically active adamantane-substituted diaza-crowns.” (PubMed, full text PDF).
    • Marjanović M et al., “Antitumor potential of crown ethers: structure-activity relationships, cell cycle disturbances, and cell death studies of a series of ionophores“, J Med Chem 2007. (PubMed, full text)

    Inferring mechanism-of-action (MOA) class of drug candidates from the differential cytostatic activity across NCI-60 cancer cell lines. Data available here >>

    • Supek F et al., “Atypical cytostatic mechanism of N-1-sulfonylcytosine derivatives determined by in vitro screening and computational analysis” Invest New Drugs 2008. (PubMed, full text)
    • Ester K et al. “Putative mechanisms of antitumor activity of cyano-substituted heteroaryles in HeLa cells” Invest New Drugs 2012. (PubMed, full text PDF)

    Antitumor activity of hydrophobic peptides bearing unnatural, adamantane-containing amino acids. Data available here >>

    • Gredičak M et al., “Computational structure-activity study directs synthesis of novel antitumor enkephalin analogs“, Amino Acids 2009. (PubMed, full text PDF)

    Site maintained by Fran Supek, RBI, Croatia.

    Data mining methods

    If you frequently perform data mining tasks on large datasets, you might find these open-source implementations of the Random Forests algorithm interesting:

    • PARF – written in Fortran 90, can parallelize its workload over a computer cluster, has many features in addition to classification (attribute importances, outlier detection, class prototypes, scaling…)
    • FastRandomForest – an efficient, multithreaded Java-based implementation that integrates seamlessly into the Weka data mining enviroment;