MCDCalc: Markov Chain Molecular Descriptors Calculator for Medicinal Chemistry.

AIMS: Cheminformatics models are able to predict different outputs (activity, property, chemical reactivity) in single molecules or complex molecular systems (catalyzed organic synthesis, metabolic reactions, nanoparticles, etc.). BACKGROUND: Cheminformatics models are able to predict different outputs (activity, property, chemical reactivity) in single molecules or complex molecular systems (catalyzed organic synthesis, metabolic reactions, nanoparticles, etc.). OBJECTIVE: Cheminformatics prediction of complex catalytic enantioselective reactions is a major goal in organic synthesis research and chemical industry. Markov Chain Molecular Descriptors (MCDs) have been largely used to solve Cheminformatics problems. There are different types of Markov chain descriptors such as Markov-Shannon entropies (Shk), Markov Means (Mk), Markov Moments (πk), etc. However, there are other possible MCDs that have not been used before. In addition, the calculation of MCDs is done very often using specific software not always available for general users and there is not an R library public available for the calculation of MCDs. This fact, limits the availability of MCMDbased Cheminformatics procedures. METHODS: We studied the enantiomeric excess ee(%)[Rcat] for 324 α-amidoalkylation reactions. These reactions have a complex mechanism depending on various factors. The model includes MCDs of the substrate, solvent, chiral catalyst, product along with values of time of reaction, temperature, load of catalyst, etc. We tested several Machine Learning regression algorithms. The Random Forest regression model has R2 > 0.90 in training and test. Secondly, the biological activity of 5644 compounds against colorectal cancer was studied. RESULTS: We developed very interesting model able to predict with Specificity and Sensitivity 70-82% the cases of preclinical assays in both training and validation series. CONCLUSION: The work shows the potential of the new tool for computational studies in organic and medicinal chemistry.

On integrating multi-experiment microarray data.

With the extensive use of microarray technology as a potential prognostic and diagnostic tool, the comparison and reproducibility of results obtained from the use of different platforms is of interest. The integration of those datasets can yield more informative results corresponding to numerous datasets and microarray platforms. We developed a novel integration technique for microarray gene-expression data derived by different studies for the purpose of a two-way Bayesian partition modelling which estimates co-expression profiles under subsets of genes and between biological samples or experimental conditions. The suggested methodology transforms disparate gene-expression data on a common probability scale to obtain inter-study-validated gene signatures. We evaluated the performance of our model using artificial data. Finally, we applied our model to six publicly available cancer gene-expression datasets and compared our results with well-known integrative microarray data methods. Our study shows that the suggested framework can relieve the limited sample size problem while reporting high accuracies by integrating multi-experiment data.

An integrated in silico approach to design specific inhibitors targeting human poly(a)-specific ribonuclease.

Poly(A)-specific ribonuclease (PARN) is an exoribonuclease/deadenylase that degrades 3'-end poly(A) tails in almost all eukaryotic organisms. Much of the biochemical and structural information on PARN comes from the human enzyme. However, the existence of PARN all along the eukaryotic evolutionary ladder requires further and thorough investigation. Although the complete structure of the full-length human PARN, as well as several aspects of the catalytic mechanism still remain elusive, many previous studies indicate that PARN can be used as potent and promising anti-cancer target. In the present study, we attempt to complement the existing structural information on PARN with in-depth bioinformatics analyses, in order to get a hologram of the molecular evolution of PARNs active site. In an effort to draw an outline, which allows specific drug design targeting PARN, an unequivocally specific platform was designed for the development of selective modulators focusing on the unique structural and catalytic features of the enzyme. Extensive phylogenetic analysis based on all the publicly available genomes indicated a broad distribution for PARN across eukaryotic species and revealed structurally important amino acids which could be assigned as potentially strong contributors to the regulation of the catalytic mechanism of PARN. Based on the above, we propose a comprehensive in silico model for the PARN's catalytic mechanism and moreover, we developed a 3D pharmacophore model, which was subsequently used for the introduction of DNP-poly(A) amphipathic substrate analog as a potential inhibitor of PARN. Indeed, biochemical analysis revealed that DNP-poly(A) inhibits PARN competitively. Our approach provides an efficient integrated platform for the rational design of pharmacophore models as well as novel modulators of PARN with therapeutic potential.

Multi-platform data integration in microarray analysis.

An increasing number of studies have profiled gene expressions in tumor specimens using distinct microarray platforms and analysis techniques. One challenging task is to develop robust statistical models in order to integrate multi-platform findings. We compare some methodologies on the field with respect to estrogen receptor (ER) status, and focus on a unified-among-platforms scale implemented by Shen et al. in 2004, which is based on a Bayesian mixture model. Under this scale, we study the ER intensity similarities between four breast cancer datasets derived from various platforms. We evaluate our results with an independent dataset in terms of ER sample classification, given the derived gene ER signatures of the integrated data. We found that integrated multi-platform gene signatures and fold-change variability similarities between different platform measurements can assist the statistical analysis of independent microarray datasets in terms of ER classification.