Intelligent resolution: Integrating Cryo-EM with AI-driven multi-resolution simulations to observe the severe acute respiratory syndrome coronavirus-2 replication-transcription machinery in action.

The severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) replication transcription complex (RTC) is a multi-domain protein responsible for replicating and transcribing the viral mRNA inside a human cell. Attacking RTC function with pharmaceutical compounds is a pathway to treating COVID-19. Conventional tools, e.g., cryo-electron microscopy and all-atom molecular dynamics (AAMD), do not provide sufficiently high resolution or timescale to capture important dynamics of this molecular machine. Consequently, we develop an innovative workflow that bridges the gap between these resolutions, using mesoscale fluctuating finite element analysis (FFEA) continuum simulations and a hierarchy of AI-methods that continually learn and infer features for maintaining consistency between AAMD and FFEA simulations. We leverage a multi-site distributed workflow manager to orchestrate AI, FFEA, and AAMD jobs, providing optimal resource utilization across HPC centers. Our study provides unprecedented access to study the SARS-CoV-2 RTC machinery, while providing general capability for AI-enabled multi-resolution simulations at scale.

Explainable AI (xAI) for Anatomic Pathology.

Pathologists are adopting whole slide images (WSIs) for diagnosis, thanks to recent FDA approval of WSI systems as class II medical devices. In response to new market forces and recent technology advances outside of pathology, a new field of computational pathology has emerged that applies artificial intelligence (AI) and machine learning algorithms to WSIs. Computational pathology has great potential for augmenting pathologists' accuracy and efficiency, but there are important concerns regarding trust of AI due to the opaque, black-box nature of most AI algorithms. In addition, there is a lack of consensus on how pathologists should incorporate computational pathology systems into their workflow. To address these concerns, building computational pathology systems with explainable AI (xAI) mechanisms is a powerful and transparent alternative to black-box AI models. xAI can reveal underlying causes for its decisions; this is intended to promote safety and reliability of AI for critical tasks such as pathology diagnosis. This article outlines xAI enabled applications in anatomic pathology workflow that improves efficiency and accuracy of the practice. In addition, we describe HistoMapr-Breast, an initial xAI enabled software application for breast core biopsies. HistoMapr-Breast automatically previews breast core WSIs and recognizes the regions of interest to rapidly present the key diagnostic areas in an interactive and explainable manner. We anticipate xAI will ultimately serve pathologists as an interactive computational guide for computer-assisted primary diagnosis.

ANCA: Anharmonic Conformational Analysis of Biomolecular Simulations.

Anharmonicity in time-dependent conformational fluctuations is noted to be a key feature of functional dynamics of biomolecules. Although anharmonic events are rare, long-timescale (µs-ms and beyond) simulations facilitate probing of such events. We have previously developed quasi-anharmonic analysis to resolve higher-order spatial correlations and characterize anharmonicity in biomolecular simulations. In this article, we have extended this toolbox to resolve higher-order temporal correlations and built a scalable Python package called anharmonic conformational analysis (ANCA). ANCA has modules to: 1) measure anharmonicity in the form of higher-order statistics and its variation as a function of time, 2) output a storyboard representation of the simulations to identify key anharmonic conformational events, and 3) identify putative anharmonic conformational substates and visualization of transitions between these substates.

Modulating Enzyme Activity by Altering Protein Dynamics with Solvent.

Optimal enzyme activity depends on a number of factors, including structure and dynamics. The role of enzyme structure is well recognized; however, the linkage between protein dynamics and enzyme activity has given rise to a contentious debate. We have developed an approach that uses an aqueous mixture of organic solvent to control the functionally relevant enzyme dynamics (without changing the structure), which in turn modulates the enzyme activity. Using this approach, we predicted that the hydride transfer reaction catalyzed by the enzyme dihydrofolate reductase (DHFR) from Escherichia coli in aqueous mixtures of isopropanol (IPA) with water will decrease by ∼3 fold at 20% (v/v) IPA concentration. Stopped-flow kinetic measurements find that the pH-independent khydride rate decreases by 2.2 fold. X-ray crystallographic enzyme structures show no noticeable differences, while computational studies indicate that the transition state and electrostatic effects were identical for water and mixed solvent conditions; quasi-elastic neutron scattering studies show that the dynamical enzyme motions are suppressed. Our approach provides a unique avenue to modulating enzyme activity through changes in enzyme dynamics. Further it provides vital insights that show the altered motions of DHFR cause significant changes in the enzyme's ability to access its functionally relevant conformational substates, explaining the decreased khydride rate. This approach has important implications for obtaining fundamental insights into the role of rate-limiting dynamics in catalysis and as well as for enzyme engineering.

Artificial intelligence techniques for integrative structural biology of intrinsically disordered proteins.

We outline recent developments in artificial intelligence (AI) and machine learning (ML) techniques for integrative structural biology of intrinsically disordered proteins (IDP) ensembles. IDPs challenge the traditional protein structure-function paradigm by adapting their conformations in response to specific binding partners leading them to mediate diverse, and often complex cellular functions such as biological signaling, self-organization and compartmentalization. Obtaining mechanistic insights into their function can therefore be challenging for traditional structural determination techniques. Often, scientists have to rely on piecemeal evidence drawn from diverse experimental techniques to characterize their functional mechanisms. Multiscale simulations can help bridge critical knowledge gaps about IDP structure-function relationships-however, these techniques also face challenges in resolving emergent phenomena within IDP conformational ensembles. We posit that scalable statistical inference techniques can effectively integrate information gleaned from multiple experimental techniques as well as from simulations, thus providing access to atomistic details of these emergent phenomena.

In situ functional cell phenotyping reveals microdomain networks in colorectal cancer recurrence.

Tumors are dynamic ecosystems comprising localized niches (microdomains), possessing distinct compositions and spatial configurations of cancer and non-cancer cell populations. Microdomain-specific network signaling coevolves with a continuum of cell states and functional plasticity associated with disease progression and therapeutic responses. We present LEAPH, an unsupervised machine learning algorithm for identifying cell phenotypes, which applies recursive steps of probabilistic clustering and spatial regularization to derive functional phenotypes (FPs) along a continuum. Combining LEAPH with pointwise mutual information and network biology analyses enables the discovery of outcome-associated microdomains visualized as distinct spatial configurations of heterogeneous FPs. Utilization of an immunofluorescence-based (51 biomarkers) image dataset of colorectal carcinoma primary tumors (n = 213) revealed microdomain-specific network dysregulation supporting cancer stem cell maintenance and immunosuppression that associated selectively with the recurrence phenotype. LEAPH enables an explainable artificial intelligence platform providing insights into pathophysiological mechanisms and novel drug targets to inform personalized therapeutic strategies.

Nucleotide substrate binding characterization in human pancreatic-type ribonucleases.

Human genome contains a group of more than a dozen similar genes with diverse biological functions including antiviral, antibacterial and angiogenesis activities. The characterized gene products of this group show significant sequence similarity and a common structural fold associated with binding and cleavage of ribonucleic acid (RNA) substrates. Therefore, these proteins have been categorized as members of human pancreatic-type ribonucleases (hRNases). hRNases differ in cell/tissue localization and display distinct substrate binding preferences and a wide range of ribonucleolytic catalytic efficiencies. Limited information is available about structural and dynamical properties that influence this diversity among these homologous RNases. Here, we use computer simulations to characterize substrate interactions, electrostatics and dynamical properties of hRNases 1-7 associated with binding to two nucleotide substrates (ACAC and AUAU). Results indicate that even with complete conservation of active-site catalytic triad associated with ribonucleolytic activity, these enzymes show significant differences in substrate interactions. Detailed characterization suggests that in addition to binding site electrostatic and van der Waals interactions, dynamics of distal regions may also play a role in binding. Another key insight is that a small difference in temperature of 300 K (used in experimental studies) and 310 K (physiological temperature) shows significant changes in enzyme-substrate interactions.

Spatial Statistics for Segmenting Histological Structures in H&E Stained Tissue Images.

Segmenting a broad class of histological structures in transmitted light and/or fluorescence-based images is a prerequisite for determining the pathological basis of cancer, elucidating spatial interactions between histological structures in tumor microenvironments (e.g., tumor infiltrating lymphocytes), facilitating precision medicine studies with deep molecular profiling, and providing an exploratory tool for pathologists. This paper focuses on segmenting histological structures in hematoxylin- and eosin-stained images of breast tissues, e.g., invasive carcinoma, carcinoma in situ, atypical and normal ducts, adipose tissue, and lymphocytes. We propose two graph-theoretic segmentation methods based on local spatial color and nuclei neighborhood statistics. For benchmarking, we curated a data set of 232 high-power field breast tissue images together with expertly annotated ground truth. To accurately model the preference for histological structures (ducts, vessels, tumor nets, adipose, etc.) over the remaining connective tissue and non-tissue areas in ground truth annotations, we propose a new region-based score for evaluating segmentation algorithms. We demonstrate the improvement of our proposed methods over the state-of-the-art algorithms in both region- and boundary-based performance measures.

Platform for Quantitative Evaluation of Spatial Intratumoral Heterogeneity in Multiplexed Fluorescence Images.

We introduce THRIVE (Tumor Heterogeneity Research Interactive Visualization Environment), an open-source tool developed to assist cancer researchers in interactive hypothesis testing. The focus of this tool is to quantify spatial intratumoral heterogeneity (ITH), and the interactions between different cell phenotypes and noncellular constituents. Specifically, we foresee applications in phenotyping cells within tumor microenvironments, recognizing tumor boundaries, identifying degrees of immune infiltration and epithelial/stromal separation, and identification of heterotypic signaling networks underlying microdomains. The THRIVE platform provides an integrated workflow for analyzing whole-slide immunofluorescence images and tissue microarrays, including algorithms for segmentation, quantification, and heterogeneity analysis. THRIVE promotes flexible deployment, a maintainable code base using open-source libraries, and an extensible framework for customizing algorithms with ease. THRIVE was designed with highly multiplexed immunofluorescence images in mind, and, by providing a platform to efficiently analyze high-dimensional immunofluorescence signals, we hope to advance these data toward mainstream adoption in cancer research. Cancer Res; 77(21); e71-74. ©2017 AACR.

Folding propensity of intrinsically disordered proteins by osmotic stress.

Proteins imparted with intrinsic disorder conduct a range of essential cellular functions. To better understand the folding and hydration properties of intrinsically disordered proteins (IDPs), we used osmotic stress to induce conformational changes in nuclear co-activator binding domain (NCBD) and activator for thyroid hormone and retinoid receptor (ACTR) separate from their mutual binding. Osmotic stress was applied by the addition of small and polymeric osmolytes, where we discovered that water contributions to NCBD folding always exceeded those for ACTR. Both NCBD and ACTR were found to gain α-helical structure with increasing osmotic stress, consistent with their folding upon NCBD/ACTR complex formation. Using small-angle neutron scattering (SANS), we further characterized NCBD structural changes with the osmolyte ethylene glycol. Here a large reduction in overall size initially occurred before substantial secondary structural change. By focusing on folding propensity, and linked hydration changes, we uncover new insights that may be important for how IDP folding contributes to binding.

Pointwise mutual information quantifies intratumor heterogeneity in tissue sections labeled with multiple fluorescent biomarkers.

BACKGROUND: Measures of spatial intratumor heterogeneity are potentially important diagnostic biomarkers for cancer progression, proliferation, and response to therapy. Spatial relationships among cells including cancer and stromal cells in the tumor microenvironment (TME) are key contributors to heterogeneity. METHODS: We demonstrate how to quantify spatial heterogeneity from immunofluorescence pathology samples, using a set of 3 basic breast cancer biomarkers as a test case. We learn a set of dominant biomarker intensity patterns and map the spatial distribution of the biomarker patterns with a network. We then describe the pairwise association statistics for each pattern within the network using pointwise mutual information (PMI) and visually represent heterogeneity with a two-dimensional map. RESULTS: We found a salient set of 8 biomarker patterns to describe cellular phenotypes from a tissue microarray cohort containing 4 different breast cancer subtypes. After computing PMI for each pair of biomarker patterns in each patient and tumor replicate, we visualize the interactions that contribute to the resulting association statistics. Then, we demonstrate the potential for using PMI as a diagnostic biomarker, by comparing PMI maps and heterogeneity scores from patients across the 4 different cancer subtypes. Estrogen receptor positive invasive lobular carcinoma patient, AL13-6, exhibited the highest heterogeneity score among those tested, while estrogen receptor negative invasive ductal carcinoma patient, AL13-14, exhibited the lowest heterogeneity score. CONCLUSIONS: This paper presents an approach for describing intratumor heterogeneity, in a quantitative fashion (via PMI), which departs from the purely qualitative approaches currently used in the clinic. PMI is generalizable to highly multiplexed/hyperplexed immunofluorescence images, as well as spatial data from complementary in situ methods including FISSEQ and CyTOF, sampling many different components within the TME. We hypothesize that PMI will uncover key spatial interactions in the TME that contribute to disease proliferation and progression.

Automated identification of abnormal respiratory ciliary motion in nasal biopsies.

Motile cilia lining the nasal and bronchial passages beat synchronously to clear mucus and foreign matter from the respiratory tract. This mucociliary defense mechanism is essential for pulmonary health, because respiratory ciliary motion defects, such as those in patients with primary ciliary dyskinesia (PCD) or congenital heart disease, can cause severe sinopulmonary disease necessitating organ transplant. The visual examination of nasal or bronchial biopsies is critical for the diagnosis of ciliary motion defects, but these analyses are highly subjective and error-prone. Although ciliary beat frequency can be computed, this metric cannot sensitively characterize ciliary motion defects. Furthermore, PCD can present without any ultrastructural defects, limiting the use of other detection methods, such as electron microscopy. Therefore, an unbiased, computational method for analyzing ciliary motion is clinically compelling. We present a computational pipeline using algorithms from computer vision and machine learning to decompose ciliary motion into quantitative elemental components. Using this framework, we constructed digital signatures for ciliary motion recognition and quantified specific properties of the ciliary motion that allowed high-throughput classification of ciliary motion as normal or abnormal. We achieved >90% classification accuracy in two independent data cohorts composed of patients with congenital heart disease, PCD, or heterotaxy, as well as healthy controls. Clinicians without specialized knowledge in machine learning or computer vision can operate this pipeline as a "black box" toolkit to evaluate ciliary motion.