Most of today’s data analysis – scientific or otherwise – is performed in a vectorized programming language (e.g., Python+numpy, R, MATLAB). These languages are distinguished from more traditional “scalar” programming languages, in that one performs operations on entire arrays, instead of looping over each individual element.

Maksims Zigunovs is a visiting PhD student from Riga Technical University in Riga, Latvia.

Traditional informatics in comparative genomics and molecular evolution work only with static representations of biomolecules (i.e. sequence and structure), thereby ignoring the molecular dynamics (MD) of proteins that defines function in the cell. A comparative approach applied to MD would connect this very short time scale process, defined in femtoseconds, to one of the longest in the universe, measured in millions of years. Here, we leverage advances in GPU-accelerated molecular dynamics simulation software to develop a comparative method of MD analysis and visualization that can be applied to any two homologous PDB structures. Our open source pipeline, DROIDS (Detecting Relative Outlier Impacts in Dynamic Simulations) works in conjunction with existing molecular modeling software to convert any Linux gaming PC into a ‘comparative computational microscope’ for observing the biophysical effects of mutations and other chemical changes in proteins. DROIDS implements structural alignment and Benjamini-Hochberg corrected Kolmogorov-Smirnov statistics to compare atom bond fluctuations (i.e. thermostability) and correlations in atom vector trajectories (i.e. concerted motions) on the protein backbone, color mapping the significant differences identified in protein MD with single amino acid resolution. DROIDS is simple to use, incorporating separate GUI controls for Amber16 MD simulation, cpptraj analysis and the final statistical and visual representations using both R graphics and UCSF Chimera. DROIDS can also potentially be utilized to visually investigate quantitative changes in molecular dynamics due to binding interactions of pharmaceuticals, toxins, or other biomolecules. Visit our code repository at https://github.com/gbabbitt/DROIDS-1.0 and for many case examples, visit our YouTube channel at https://www.youtube.com/channel/UCJTBqGq01pBCMDQikn566Kw

This is the first meeting of the semester for our group.

With the advent of gravitational wave (GW) astronomy, following the first GW detection in September 2015 by the LIGO Scientific Collaboration and the Virgo Collaboration, we now have a means to measure the properties of GW emitters, including merging compact objects such as black holes and neutron stars. The expected number of merger events in a given volume of space and observation time can be modeled as an inhomogeneous Poisson process, with an overall rate and a probability distribution over the systems’ various physical properties. This rate and distribution will provide information on the population of binary star systems, and can be compared with theoretical population models. In this talk, we describe a general Bayesian framework for inferring this rate and distribution, using Markov chain Monte Carlo (MCMC). This method works on noisy data, and relies on a measure of the volume in space to which our detectors are sensitive (which is itself a function of the physical properties). This method relies on a parameterization of the probability distribution of system parameters, which can be restrictive (e.g., a power law) or flexible (e.g., a Gaussian mixture model). With some slight modifications, this method may be used in a variety of terrestrial applications.