Searcharxiv⌕ Search

arXiv subjects

Cameron Mura

Publications and source records attributed to Cameron Mura.

At least 19 recordsLinked to original sources
arXiv2026

Phil Bourne (1953-2026): From Small Molecules to Big Data --- The Journey of a Multifaceted Visionary

Born in London in 1953 and raised in Australia, Phil Bourne spent over four decades in science moving across scales: from crystal structures to the world's premier structural biology database; from scientific journals to national data policy; from molecules to institutions. The International Society for Computational Biology (ISCB) chose Phil to receive its 2026 Outstanding Service Award (Wiper, 2026). He passed away on 8 March 2026, in Charlottesville, Virginia, before he could receive it. This piece is in memory of Phil, who most recently served as the Founding Stephenson Dean of the University of Virginia's School of Data Science and as a Professor of Biomedical Engineering$\unicode{x2013}$and who was, over the course of his career, a steadfast builder of Computational Biology and Bioinformatics, of Open Science and Data Science. A tribute site, https://philbourne.muchloved.com, invites those who knew Phil to share memories, photos, and reflections.

q-bio.OT↗
arXiv2022

A Tribute to Phil Bourne -- Scientist and Human

This Special Issue of Biomolecules, commissioned in honor of Dr. Philip E. Bourne, focuses on a new field of biomolecular data science. In this brief retrospective, we consider the arc of Bourne's 40-year scientific and professional career, particularly as it relates to the origins of this new field.

q-bio.OT↗
arXiv2022

Innovations in Integrating Machine Learning and Agent-Based Modeling of Biomedical Systems

Agent-based modeling (ABM) is a well-established paradigm for simulating complex systems via interactions between constituent entities. Machine learning (ML) refers to approaches whereby statistical algorithms 'learn' from data on their own, without imposing a priori theories of system behavior. Biological systems -- from molecules, to cells, to entire organisms -- consist of vast numbers of entities, governed by complex webs of interactions that span many spatiotemporal scales and exhibit nonlinearity, stochasticity and intricate coupling between entities. The macroscopic properties and collective dynamics of such systems are difficult to capture via continuum modelling and mean-field formalisms. ABM takes a 'bottom-up' approach that obviates these difficulties by enabling one to easily propose and test a set of well-defined 'rules' to be applied to the individual entities (agents) in a system. Evaluating a system and propagating its state over discrete time-steps effectively simulates the system, allowing observables to be computed and system properties to be analyzed. Because the rules that govern an ABM can be difficult to abstract and formulate from experimental data, there is an opportunity to use ML to help infer optimal, system-specific ABM rules. Once such rule-sets are devised, ABM calculations can generate a wealth of data, and ML can be applied there too -- e.g., to probe statistical measures that meaningfully describe a system's stochastic properties. As an example of synergy in the other direction (from ABM to ML), ABM simulations can generate realistic datasets for training ML algorithms (e.g., for regularization, to mitigate overfitting). In these ways, one can envision various synergistic ABM$\rightleftharpoons$ML loops. This review summarizes how ABM and ML have been integrated in contexts that span spatiotemporal scales, from cellular to population-level epidemiology.

q-bio.QM↗
arXiv2021

Exploration of Dark Chemical Genomics Space via Portal Learning: Applied to Targeting the Undruggable Genome and COVID-19 Anti-Infective Polypharmacology

Advances in biomedicine are largely fueled by exploring uncharted territories of human biology. Machine learning can both enable and accelerate discovery, but faces a fundamental hurdle when applied to unseen data with distributions that differ from previously observed ones -- a common dilemma in scientific inquiry. We have developed a new deep learning framework, called {\textit{Portal Learning}}, to explore dark chemical and biological space. Three key, novel components of our approach include: (i) end-to-end, step-wise transfer learning, in recognition of biology's sequence-structure-function paradigm, (ii) out-of-cluster meta-learning, and (iii) stress model selection. Portal Learning provides a practical solution to the out-of-distribution (OOD) problem in statistical machine learning. Here, we have implemented Portal Learning to predict chemical-protein interactions on a genome-wide scale. Systematic studies demonstrate that Portal Learning can effectively assign ligands to unexplored gene families (unknown functions), versus existing state-of-the-art methods, thereby allowing us to target previously "undruggable" proteins and design novel polypharmacological agents for disrupting interactions between SARS-CoV-2 and human proteins. Portal Learning is general-purpose and can be further applied to other areas of scientific inquiry.

q-bio.QM↗
arXiv2021

A Birds-eye (Re)View of Acid-suppression Drugs, COVID-19, and the Highly Variable Literature

We consider the recent surge of information on the potential benefits of acid-suppression drugs in the context of COVID-19, with an eye on the variability (and confusion) across the reported findings--at least as regards the popular antacid famotidine. The inconsistencies reflect contradictory conclusions from independent clinical-based studies that took roughly similar approaches, in terms of experimental design (retrospective, cohort-based, etc.) and statistical analyses (propensity-score matching and stratification, etc.). The confusion has significant ramifications in choosing therapeutic interventions: e.g., do potential benefits of famotidine indicate its use in a particular COVID-19 case? Beyond this pressing therapeutic issue, conflicting information on famotidine must be resolved before its integration in ontological and knowledge graph-based frameworks, which in turn are useful in drug repurposing efforts. To begin systematically structuring the rapidly accumulating information, in the hopes of clarifying and reconciling the discrepancies, we consider the contradictory information along three proposed 'axes': (1) a context-of-disease axis, (2) a degree-of-[therapeutic]-benefit axis, and (3) a mechanism-of-action axis. We suspect that incongruencies in how these axes have been (implicitly) treated in past studies has led to the contradictory indications for famotidine and COVID-19. We also trace the evolution of information on acid-suppression agents as regards the transmission, severity, and mortality of COVID-19, given the many literature reports that have accumulated. By grouping the studies conceptually and thematically, we identify three eras in the progression of our understanding of famotidine and COVID-19. Harmonizing these findings is a key goal for both clinical standards-of-care (COVID and beyond) as well as ontological and knowledge graph-based approaches.

q-bio.TO↗
arXiv2020

Deep Learning of Protein Structural Classes: Any Evidence for an 'Urfold'?

Recent computational advances in the accurate prediction of protein three-dimensional (3D) structures from amino acid sequences now present a unique opportunity to decipher the interrelationships between proteins. This task entails--but is not equivalent to--a problem of 3D structure comparison and classification. Historically, protein domain classification has been a largely manual and subjective activity, relying upon various heuristics. Databases such as CATH represent significant steps towards a more systematic (and automatable) approach, yet there still remains much room for the development of more scalable and quantitative classification methods, grounded in machine learning. We suspect that re-examining these relationships via a Deep Learning (DL) approach may entail a large-scale restructuring of classification schemes, improved with respect to the interpretability of distant relationships between proteins. Here, we describe our training of DL models on protein domain structures (and their associated physicochemical properties) in order to evaluate classification properties at CATH's "homologous superfamily" (SF) level. To achieve this, we have devised and applied an extension of image-classification methods and image segmentation techniques, utilizing a convolutional autoencoder model architecture. Our DL architecture allows models to learn structural features that, in a sense, 'define' different homologous SFs. We evaluate and quantify pairwise 'distances' between SFs by building one model per SF and comparing the loss functions of the models. Hierarchical clustering on these distance matrices provides a new view of protein interrelationships--a view that extends beyond simple structural/geometric similarity, and towards the realm of structure/function properties.

q-bio.BM↗
arXiv2019

Machine Learning for Classification of Protein Helix Capping Motifs

The biological function of a protein stems from its 3-dimensional structure, which is thermodynamically determined by the energetics of interatomic forces between its amino acid building blocks (the order of amino acids, known as the sequence, defines a protein). Given the costs (time, money, human resources) of determining protein structures via experimental means such as X-ray crystallography, can we better describe and compare protein 3D structures in a robust and efficient manner, so as to gain meaningful biological insights? We begin by considering a relatively simple problem, limiting ourselves to just protein secondary structural elements. Historically, many computational methods have been devised to classify amino acid residues in a protein chain into one of several discrete secondary structures, of which the most well-characterized are the geometrically regular $α$-helix and $β$-sheet; irregular structural patterns, such as 'turns' and 'loops', are less understood. Here, we present a study of Deep Learning techniques to classify the loop-like end cap structures which delimit $α$-helices. Previous work used highly empirical and heuristic methods to manually classify helix capping motifs. Instead, we use structural data directly--including (i) backbone torsion angles computed from 3D structures, (ii) macromolecular feature sets (e.g., physicochemical properties), and (iii) helix cap classification data (from CAPS-DB)--as the ground truth to train a bidirectional long short-term memory (BiLSTM) model to classify helix cap residues. We tried different network architectures and scanned hyperparameters in order to train and assess several models; we also trained a Support Vector Classifier (SVC) to use as a baseline. Ultimately, we achieved 85% class-balanced accuracy with a deep BiLSTM model.

q-bio.BM↗
arXiv2019

Ten Quick Tips for Using a Raspberry Pi

Much of biology (and, indeed, all of science) is becoming increasingly computational. We tend to think of this in regards to algorithmic approaches and software tools, as well as increased computing power. There has also been a shift towards slicker, packaged solutions--which mirrors everyday life, from smart phones to smart homes. As a result, it's all too easy to be detached from the fundamental elements that power these changes, and to see solutions as "black boxes". The major goal of this piece is to use the example of the Raspberry Pi--a small, general-purpose computer--as the central component in a highly developed ecosystem that brings together elements like external hardware, sensors and controllers, state-of-the-art programming practices, and basic electronics and physics, all in an approachable and useful way. External devices and inputs are easily connected to the Pi, and it can, in turn, control attached devices very simply. So whether you want to use it to manage laboratory equipment, sample the environment, teach bioinformatics, control your home security or make a model lunar lander, it's all built from the same basic principles. To quote Richard Feynman, "What I cannot create, I do not understand".

cs.CY↗
arXiv2018

Structural biology meets data science: Does anything change?

Data science has emerged from the proliferation of digital data, coupled with advances in algorithms, software and hardware (e.g., GPU computing). Innovations in structural biology have been driven by similar factors, spurring us to ask: can these two fields impact one another in deep and hitherto unforeseen ways? We posit that the answer is yes. New biological knowledge lies in the relationships between sequence, structure, function and disease, all of which play out on the stage of evolution, and data science enables us to elucidate these relationships at scale. Here, we consider the above question from the five key pillars of data science: acquisition, engineering, analytics, visualization and policy, with an emphasis on machine learning as the premier analytics approach.

q-bio.QM↗
arXiv2016

Toward a Designable Extracellular Matrix: Molecular Dynamics Simulations of an Engineered Laminin-mimetic, Elastin-like Fusion Protein

Native extracellular matrices (ECMs), such as those of the human brain and other neural tissues, exhibit networks of molecular interactions between specific matrix proteins and other tissue components. Guided by these naturally self-assembling supramolecular systems, we have designed a matrix-derived protein chimera that contains a laminin globular-like (LG) domain fused to an elastin-like polypeptide (ELP). All-atom, classical molecular dynamics simulations of our designed laminin-elastin fusion protein reveal temperature-dependent conformational changes, in terms of secondary structure composition, solvent accessible surface area, hydrogen bonding, and surface hydration. These properties illuminate the phase behavior of this fusion protein, via the emergence of $β$-sheet character in physiologically-relevant temperature ranges.

q-bio.BM↗
arXiv2016

The Structures, Functions, and Evolution of Sm-like Archaeal Proteins (SmAPs)

Sm proteins were discovered nearly 20 years ago as a group of small antigenic proteins ($\approx$ 90-120 residues). Since then, an extensive amount of biochemical and genetic data have illuminated the crucial roles of these proteins in forming ribonucleoprotein (RNP) complexes that are used in RNA processing, e.g., spliceosomal removal of introns from pre-mRNAs. Spliceosomes are large macromolecular machines that are comparable to ribosomes in size and complexity, and are composed of uridine-rich small nuclear RNPs (U snRNPs). Various sets of seven different Sm proteins form the cores of most snRNPs. Despite their importance, very little is known about the atomic-resolution structure of snRNPs or their Sm cores. As a first step towards a high-resolution image of snRNPs and their hierarchic assembly, we have determined the crystal structures of archaeal homologs of Sm proteins, which we term Sm-like archaeal proteins (SmAPs).

q-bio.BM↗
arXiv2016

Claws, Disorder, and Conformational Dynamics of the C-terminal Region of Human Desmoplakin

Multicellular organisms consist of cells that interact via elaborate adhesion complexes. Desmosomes are membrane-associated adhesion complexes that mechanically tether the cytoskeletal intermediate filaments (IFs) between two adjacent cells, creating a network of tough connections in tissues such as skin and heart. Desmoplakin (DP) is the key desmosomal protein that binds IFs, and the DP-IF association poses a quandary: desmoplakin must stably and tightly bind IFs to maintain the structural integrity of the desmosome. Yet, newly synthesized DP must traffick along the cytoskeleton to the site of nascent desmosome assembly without 'sticking' to the IF network, implying weak or transient DP--IF contacts. Recent work reveals that these contacts are modulated by post-translational modifications (PTMs) in DP's C-terminal tail. Using molecular dynamics simulations, we have elucidated the structural basis of these PTM-induced effects. Our simulations, nearing 2 microseconds in aggregate, indicate that phosphorylation of S2849 induces an 'arginine claw' in desmoplakin's C-terminal tail (DPCTT). If a key arginine, R2834, is methylated, the DPCTT preferentially samples conformations that are geometrically well-suited as substrates for processive phosphorylation by the cognate kinase GSK3. We suggest that DPCTT is a molecular switch that modulates, via its conformational dynamics, DP's efficacy as a substrate for GSK3. Finally, we show that the fluctuating DPCTT can contact other parts of DP, suggesting a competitive binding mechanism for the modulation of DP--IF interactions.

q-bio.BM↗
arXiv2016

An Introduction to Programming for Bioscientists: A Python-based Primer

Computing has revolutionized the biological sciences over the past several decades, such that virtually all contemporary research in the biosciences utilizes computer programs. The computational advances have come on many fronts, spurred by fundamental developments in hardware, software, and algorithms. These advances have influenced, and even engendered, a phenomenal array of bioscience fields, including molecular evolution and bioinformatics; genome-, proteome-, transcriptome- and metabolome-wide experimental studies; structural genomics; and atomistic simulations of cellular-scale molecular assemblies as large as ribosomes and intact viruses. In short, much of post-genomic biology is increasingly becoming a form of computational biology. The ability to design and write computer programs is among the most indispensable skills that a modern researcher can cultivate. Python has become a popular programming language in the biosciences, largely because (i) its straightforward semantics and clean syntax make it a readily accessible first language; (ii) it is expressive and well-suited to object-oriented programming, as well as other modern paradigms; and (iii) the many available libraries and third-party toolkits extend the functionality of the core language into virtually every biological domain (sequence and structure analyses, phylogenomics, workflow management systems, etc.). This primer offers a basic introduction to coding, via Python, and it includes concrete examples and exercises to illustrate the language's usage and capabilities; the main text culminates with a final project in structural bioinformatics. A suite of Supplemental Chapters is also provided. Starting with basic concepts, such as that of a 'variable', the Chapters methodically advance the reader to the point of writing a graphical user interface to compute the Hamming distance between two DNA sequences.

q-bio.OT↗
arXiv2015

Known Structure, Unknown Function: An Inquiry-based Undergraduate Biochemistry Laboratory Course

Undergraduate biochemistry laboratory courses often do not provide students with an authentic research experience, particularly when the express purpose of the laboratory is purely instructional. However, an instructional laboratory course that is inquiry- and research-based could simultaneously impart scientific knowledge and foster a student's research expertise and confidence. We have developed a year-long undergraduate biochemistry laboratory curriculum wherein students determine, via experiment and computation, the function of a protein of known three-dimensional structure. The first half of the course is inquiry-based and modular in design; students learn general biochemical techniques while gaining preparation for research experiments in the second semester. Having learned standard biochemical methods in the first semester, students independently pursue their own (original) research projects in the second semester. This new curriculum has yielded an improvement in student performance and confidence as assessed by various metrics. To disseminate teaching resources to students and instructors alike, a freely accessible Biochemistry Laboratory Education resource is available at http://biochemlab.org.

physics.ed-ph↗
arXiv2014

Development & Implementation of a PyMOL 'putty' Representation

The PyMOL molecular graphics program has been modified to introduce a new 'putty' cartoon representation, akin to the 'sausage'-style representation of the MOLMOL molecular visualization (MolVis) software package. This document outlines the development and implementation of the putty representation.

q-bio.BM↗
arXiv2014

Abstractions, Algorithms and Data Structures for Structural Bioinformatics in PyCogent

To facilitate flexible and efficient structural bioinformatics analyses, new functionality for three-dimensional structure processing and analysis has been introduced into PyCogent -- a popular feature-rich framework for sequence-based bioinformatics, but one which has lacked equally powerful tools for handling stuctural/coordinate-based data. Extensible Python modules have been developed, which provide object-oriented abstractions (based on a hierarchical representation of macromolecules), efficient data structures (e.g. kD-trees), fast implementations of common algorithms (e.g. surface-area calculations), read/write support for Protein Data Bank-related file formats and wrappers for external command-line applications (e.g. Stride). Integration of this code into PyCogent is symbiotic, allowing sequence-based work to benefit from structure-derived data and, reciprocally, enabling structural studies to leverage PyCogent's versatile tools for phylogenetic and evolutionary analyses.

q-bio.BM↗
arXiv2014

Molecular Dynamics of a kB DNA Element: Base Flipping via Cross-strand Intercalative Stacking in a Microsecond-scale Simulation

The sequence-dependent structural variability and conformational dynamics of DNA play pivotal roles in many biological milieus, such as in the site-specific binding of transcription factors to target regulatory elements. To better understand DNA structure, function, and dynamics in general, and protein-DNA recognition in the 'kB' family of genetic regulatory elements in particular, we performed molecular dynamics simulations of a 20-base pair DNA encompassing a cognate kB site recognized by the proto-oncogenic 'c-Rel' subfamily of NF-kB transcription factors. Simulations of the kB DNA in explicit water were extended to microsecond duration, providing a broad, atomically-detailed glimpse into the structural and dynamical behavior of double helical DNA over many timescales. Of particular note, novel (and structurally plausible) conformations of DNA developed only at the long times sampled in this simulation -- including a peculiar state arising at ~ 0.7 us and characterized by cross-strand intercalative stacking of nucleotides within a longitudinally-sheared base pair, followed (at ~ 1 us) by spontaneous base flipping of a neighboring thymine within the A-rich duplex. Results and predictions from the us-scale simulation include implications for a dynamical NF-kB recognition motif, and are amenable to testing and further exploration via specific experimental approaches that are suggested herein.

q-bio.BM↗