Biomarker selection and classification of "-omics" data using a two-step bayes classification framework.

Identification of suitable biomarkers for accurate prediction of phenotypic outcomes is a goal for personalized medicine. However, current machine learning approaches are either too complex or perform poorly. Here, a novel two-step machine-learning framework is presented to address this need. First, a Naïve Bayes estimator is used to rank features from which the top-ranked will most likely contain the most informative features for prediction of the underlying biological classes. The top-ranked features are then used in a Hidden Naïve Bayes classifier to construct a classification prediction model from these filtered attributes. In order to obtain the minimum set of the most informative biomarkers, the bottom-ranked features are successively removed from the Naïve Bayes-filtered feature list one at a time, and the classification accuracy of the Hidden Naïve Bayes classifier is checked for each pruned feature set. The performance of the proposed two-step Bayes classification framework was tested on different types of -omics datasets including gene expression microarray, single nucleotide polymorphism microarray (SNParray), and surface-enhanced laser desorption/ionization time-of-flight (SELDI-TOF) proteomic data. The proposed two-step Bayes classification framework was equal to and, in some cases, outperformed other classification methods in terms of prediction accuracy, minimum number of classification markers, and computational time.

Dynamics and mechanism of structural diffusion in linear hydrogen bond.

Dynamics and mechanism of proton transfer in a protonated hydrogen bond (H-bond) chain were studied, using the CH(3)OH(2)(+)(CH(3)OH)(n) complexes, n = 1-4, as model systems. The present investigations used B3LYP/TZVP calculations and Born-Oppenheimer MD (BOMD) simulations at 350 K to obtain characteristic H-bond structures, energetic and IR spectra of the transferring protons in the gas phase and continuum liquid. The static and dynamic results were compared with the H(3)O(+)(H(2)O)(n) and CH(3)OH(2)(+)(H(2)O)(n) complexes, n = 1-4. It was found that the H-bond chains with n = 1 and 3 represent the most active intermediate states and the CH(3)OH(2)(+)(CH(3)OH)(n) complexes possess the lowest threshold frequency of proton transfer. The IR spectra obtained from BOMD simulations revealed that the thermal energy fluctuation and dynamics help promote proton transfer in the shared-proton structure with n = 3 by lowering the vibrational energy for the interconversion between the oscillatory shuttling and structural diffusion motions, leading to a higher population of the structural diffusion motion than in the shared-proton structure with n = 1. Additional explanation on the previously proposed mechanisms was introduced, with the emphases on the energetic of the transferring proton, the fluctuation of the number of the CH(3)OH molecules in the H-bond chain, and the quasi-dynamic equilibriums between the shared-proton structure (n = 3) and the close-contact structures (n ≥ 4). The latter prohibits proton transfer reaction in the H-bond chain from being concerted, since the rate of the structural diffusion depends upon the lifetime of the shared-proton intermediate state.

Proton transfer reactions and dynamics in CH(3)OH-H(3)O(+)-H(2)O complexes.

Proton transfer reactions and dynamics in hydrated complexes formed from CH(3)OH, H(3)O(+) and H(2)O were studied using theoretical methods. The investigations began with searching for equilibrium structures at low hydration levels using the DFT method, from which active H-bonds in the gas phase and continuum aqueous solution were characterized and analyzed. Based on the asymmetric stretching coordinates (Deltad(DA)), four H-bond complexes were identified as potential transition states, in which the most active unit is represented by an excess proton nearly equally shared between CH(3)OH and H(2)O. These cannot be definitive due to the lack of asymmetric O-H stretching frequencies (nu(OH)) which are spectral signatures of transferring protons. Born-Oppenheimer molecular dynamics (BOMD) simulations revealed that, when the thermal energy fluctuations and dynamics were included in the model calculations, the spectral signatures at nu(OH) approximately 1000 cm(-1) appeared. In continuum aqueous solution, the H-bond complex with incomplete water coordination at charged species turned out to be the only active transition state. Based on the assumption that the thermal energy fluctuations and dynamics could temporarily break the H-bonds linking the transition state complex and water molecules in the second hydration shell, elementary reactions of proton transfer were proposed. The present study showed that, due to the coupling among various vibrational modes, the discussions on proton transfer reactions cannot be made based solely on static proton transfer potentials. Inclusion of thermal energy fluctuations and dynamics in the model calculations, as in the case of BOMD simulations, together with systematic IR spectral analyses, have been proved to be the most appropriate theoretical approaches.