DROIDS 3.0—Detecting Genetic and Drug Class Variant Impact on Conserved Protein Binding Dynamics

Abstract

The application of statistical methods to comparatively framed questions about the molecular dynamics (MD) of proteins can potentially enable investigations of biomolecular function beyond the current sequence and structural methods in bioinformatics. However, the chaotic behavior in single MD trajectories requires statistical inference that is derived from large ensembles of simulations representing the comparative functional states of a protein under investigation. Meaningful interpretation of such complex forms of big data poses serious challenges to users of MD. Here, we announce Detecting Relative Outlier Impacts from Molecular Dynamic Simulation (DROIDS) 3.0, a method and software package for comparative protein dynamics that includes maxDemon 1.0, a multimethod machine learning application that trains on large ensemble comparisons of concerted protein motions in opposing functional states generated by DROIDS and deploys learned classifications of these states onto newly generated MD simulations. Local canonical correlations in learning patterns generated from independent, yet identically prepared, MD validation runs are used to identify regions of functionally conserved protein dynamics. The subsequent impacts of genetic and/or drug class variants on conserved dynamics can also be analyzed by deploying the classifiers on variant MD simulations and quantifying how often these altered protein systems display opposing functional states. Here, we present several case studies of complex changes in functional protein dynamics caused by temperature, genetic mutation, and binding interactions with nucleic acids and small molecules. We demonstrate that our machine learning algorithm can properly identify regions of functionally conserved dynamics in ubiquitin and TATA-binding protein (TBP). We quantify the impact of genetic variation in TBP and drug class variation targeting the ATP-binding region of Hsp90 on conserved dynamics. We identify regions of conserved dynamics in Hsp90 that connect the ATP binding pocket to other functional regions. We also demonstrate that dynamic impacts of various Hsp90 inhibitors rank accordingly with how closely they mimic natural ATP binding.

Information theoretics for the machine learning detection of functionally conserved and coordinated protein motions

“…We outline a general protocol for combining molecular dynamics simulation software with machine learning packages for general programming languages with a goal of detecting functionally conserved protein dynamics related to functional binding interactions. Note: this is also the protocol implemented in our most recent software release (15). Throughout the rest of the article we abbreviate this protocol as FCDA (Functionally Conserved Dynamic Analysis) A) Utilizing one of several popular molecular dynamic (MD) simulation software suites (e.g.…”

“…These data should be extracted over many time slices to compose a univariate or multivariate feature vector for the chosen machine learning classification model. See equation 4 in (15). C) At single amino acid resolution, train one or more machine learning classifiers implemented in R or Python on these features (i.e.…”

“…For comparison, we compared this protein-protein interaction to a DNA-protein interaction in the form of human TATA binding protein (TBP) bound to DNA (33). The software implementation of the FCDA protocol is recently published by our group (15) and software pipeline was named 'maxDemon' in reference to Maxwell's mythological demon that observes the motion of atoms to achieve information gain.…”

“…We similarly obtained human TBP structure (PDB: 1cdw). Large ensembles of graphic processing unit (GPU) accelerated molecular dynamic simulations were prepared and conducted using the particle mesh Ewald method employed by pmemd.cuda in Amber18 (11,34) via the DROIDS v3.8 interface (Detecting Relative Outlier Impacts in Dynamic Simulation) (15,16). Simulations were run on a Linux Mint 19 operating system mounting two Nvidia Titan Xp graphics processors.…”

“…accurate) molecular dynamic simulations (10,11), and now greatly enhance our ability to investigate the link between molecular dynamics and evolutionarily-conserved function (9,(12)(13)(14). However, very few methods and software currently exist allowing the exploration of function and evolution through molecular dynamics simulation (15,16). Here, we propose a generalizable method of employing machine learning-based classification of functional dynamics under the same information theoretic framework that has proven so useful for the comparative analysis of multiple-aligned protein and DNA sequences (7,8).…”