Transferring a Molecular Foundation Model for Polymer Property Predictions.

Transformer-based large language models have remarkable potential to accelerate design optimization for applications such as drug development and material discovery. Self-supervised pretraining of transformer models requires large-scale data sets, which are often sparsely populated in topical areas such as polymer science. State-of-the-art approaches for polymers conduct data augmentation to generate additional samples but unavoidably incur extra computational costs. In contrast, large-scale open-source data sets are available for small molecules and provide a potential solution to data scarcity through transfer learning. In this work, we show that using transformers pretrained on small molecules and fine-tuned on polymer properties achieves comparable accuracy to those trained on augmented polymer data sets for a series of benchmark prediction tasks.

Effect of constitutive law on the erythrocyte membrane response to large strains.

Three constitutive laws, that is the Skalak, neo-Hookean and Yeoh laws, commonly employed for describing the erythrocyte membrane mechanics are theoretically analyzed and numerically investigated to assess their accuracy for capturing erythrocyte deformation characteristics and morphology. Particular emphasis is given to the nonlinear deformation regime, where it is known that the discrepancies between constitutive laws are most prominent. Hence, the experiments of optical tweezers and micropipette aspiration are considered here, for which relationships between the individual shear elastic moduli of the constitutive laws can also be established through analysis of the tension-deformation relationship. All constitutive laws were found to adequately predict the axial and transverse deformations of a red blood cell subjected to stretching with optical tweezers for a constant shear elastic modulus value. As opposed to Skalak law, the neo-Hookean and Yeoh laws replicated the erythrocyte membrane folding, that has been experimentally observed, with the trade-off of sustaining significant area variations. For the micropipette aspiration, the suction pressure-aspiration length relationship could be excellently predicted for a fixed shear elastic modulus value only when Yeoh law was considered. Importantly, the neo-Hookean and Yeoh laws reproduced the membrane wrinkling at suction pressures close to those experimentally measured. None of the constitutive laws suffered from membrane area compressibility in the micropipette aspiration case.

Propagation pattern for moment representation of the lattice Boltzmann method.

A propagation pattern for the moment representation of the regularized lattice Boltzmann method (LBM) in three dimensions is presented. Using effectively lossless compression, the simulation state is stored as a set of moments of the lattice Boltzmann distribution function, instead of the distribution function itself. An efficient cache-aware propagation pattern for this moment representation has the effect of substantially reducing both the storage and memory bandwidth required for LBM simulations. This paper extends recent work with the moment representation by expanding the performance analysis on central processing unit (CPU) architectures, considering how boundary conditions are implemented, and demonstrating the effectiveness of the moment representation on a graphics processing unit (GPU) architecture.

Language models for the prediction of SARS-CoV-2 inhibitors.

The COVID-19 pandemic highlights the need for computational tools to automate and accelerate drug design for novel protein targets. We leverage deep learning language models to generate and score drug candidates based on predicted protein binding affinity. We pre-trained a deep learning language model (BERT) on ∼9.6 billion molecules and achieved peak performance of 603 petaflops in mixed precision. Our work reduces pre-training time from days to hours, compared to previous efforts with this architecture, while also increasing the dataset size by nearly an order of magnitude. For scoring, we fine-tuned the language model using an assembled set of thousands of protein targets with binding affinity data and searched for inhibitors of specific protein targets, SARS-CoV-2 Mpro and PLpro. We utilized a genetic algorithm approach for finding optimal candidates using the generation and scoring capabilities of the language model. Our generalizable models accelerate the identification of inhibitors for emerging therapeutic targets.

Accelerated training of bootstrap aggregation-based deep information extraction systems from cancer pathology reports.

OBJECTIVE: In machine learning, it is evident that the classification of the task performance increases if bootstrap aggregation (bagging) is applied. However, the bagging of deep neural networks takes tremendous amounts of computational resources and training time. The research question that we aimed to answer in this research is whether we could achieve higher task performance scores and accelerate the training by dividing a problem into sub-problems. MATERIALS AND METHODS: The data used in this study consist of free text from electronic cancer pathology reports. We applied bagging and partitioned data training using Multi-Task Convolutional Neural Network (MT-CNN) and Multi-Task Hierarchical Convolutional Attention Network (MT-HCAN) classifiers. We split a big problem into 20 sub-problems, resampled the training cases 2,000 times, and trained the deep learning model for each bootstrap sample and each sub-problem-thus, generating up to 40,000 models. We performed the training of many models concurrently in a high-performance computing environment at Oak Ridge National Laboratory (ORNL). RESULTS: We demonstrated that aggregation of the models improves task performance compared with the single-model approach, which is consistent with other research studies; and we demonstrated that the two proposed partitioned bagging methods achieved higher classification accuracy scores on four tasks. Notably, the improvements were significant for the extraction of cancer histology data, which had more than 500 class labels in the task; these results show that data partition may alleviate the complexity of the task. On the contrary, the methods did not achieve superior scores for the tasks of site and subsite classification. Intrinsically, since data partitioning was based on the primary cancer site, the accuracy depended on the determination of the partitions, which needs further investigation and improvement. CONCLUSION: Results in this research demonstrate that 1. The data partitioning and bagging strategy achieved higher performance scores. 2. We achieved faster training leveraged by the high-performance Summit supercomputer at ORNL.

Hemodynamic and morphological characteristics of a growing cerebral aneurysm.

The growth of cerebral aneurysms is linked to local hemodynamic conditions, but the driving mechanisms of the growth are poorly understood. The goal of this study was to examine the association between intraaneurysmal hemodynamic features and areas of aneurysm growth, to present the key hemodynamic parameters essential for an accurate prediction of the growth, and to gain a deeper understanding of the underlying mechanisms. Patient-specific images of a growing cerebral aneurysm in 3 different growth stages acquired over a period of 40 months were segmented and reconstructed. A unique aspect of this patient-specific case study was that while one side of the aneurysm stayed stable, the other side continued to grow. This unique case enabled the authors to examine their aims in the same patient with parent and daughter arteries under the same inlet flow conditions. Pulsatile flow in the aneurysm models was simulated using computational fluid dynamics and was validated with in vitro experiments using particle image velocimetry measurements. The authors' detailed analysis of intrasaccular hemodynamics linked the growing regions of aneurysms to flow instabilities and complex vortex structures. Extremely low velocities were observed at or around the center of the unstable vortex structure, which matched well with the growing regions of the studied cerebral aneurysm. Furthermore, the authors observed that the aneurysm wall regions with a growth greater than 0.5 mm coincided with wall regions of lower (< 0.5 Pa) time-averaged wall shear stress (TAWSS), lower instantaneous (< 0.5 Pa) wall shear stress (WSS), and high (> 0.1) oscillatory shear index (OSI). To determine which set of parameters can best identify growing and nongrowing aneurysms, the authors performed statistical analysis for consecutive stages of the growing CA. The results demonstrated that the combination of TAWSS and the distance from the center of the vortical structure has the highest sensitivity and positive predictive value, and relatively high specificity and negative predictive value. These findings suggest that an unstable, recirculating flow structure within the aneurysm sac created in the region adjacent to the aneurysm wall with low TAWSS may be introduced as an accurate criterion to explain the hemodynamic conditions predisposing the aneurysm to growth. The authors' findings are based on one patient's data set, but the study lays out the justification for future large-scale verification. The authors' findings can assist clinicians in differentiating stable and growing aneurysms during preinterventional planning.

Enhancing molecular design efficiency: Uniting language models and generative networks with genetic algorithms.

This study examines the effectiveness of generative models in drug discovery, material science, and polymer science, aiming to overcome constraints associated with traditional inverse design methods relying on heuristic rules. Generative models generate synthetic data resembling real data, enabling deep learning model training without extensive labeled datasets. They prove valuable in creating virtual libraries of molecules for material science and facilitating drug discovery by generating molecules with specific properties. While generative adversarial networks (GANs) are explored for these purposes, mode collapse restricts their efficacy, limiting novel structure variability. To address this, we introduce a masked language model (LM) inspired by natural language processing. Although LMs alone can have inherent limitations, we propose a hybrid architecture combining LMs and GANs to efficiently generate new molecules, demonstrating superior performance over standalone masked LMs, particularly for smaller population sizes. This hybrid LM-GAN architecture enhances efficiency in optimizing properties and generating novel samples.

Path-BigBird: An AI-Driven Transformer Approach to Classification of Cancer Pathology Reports.

PURPOSE: Surgical pathology reports are critical for cancer diagnosis and management. To accurately extract information about tumor characteristics from pathology reports in near real time, we explore the impact of using domain-specific transformer models that understand cancer pathology reports. METHODS: We built a pathology transformer model, Path-BigBird, by using 2.7 million pathology reports from six SEER cancer registries. We then compare different variations of Path-BigBird with two less computationally intensive methods: Hierarchical Self-Attention Network (HiSAN) classification model and an off-the-shelf clinical transformer model (Clinical BigBird). We use five pathology information extraction tasks for evaluation: site, subsite, laterality, histology, and behavior. Model performance is evaluated by using macro and micro F1 scores. RESULTS: We found that Path-BigBird and Clinical BigBird outperformed the HiSAN in all tasks. Clinical BigBird performed better on the site and laterality tasks. Versions of the Path-BigBird model performed best on the two most difficult tasks: subsite (micro F1 score of 72.53, macro F1 score of 35.76) and histology (micro F1 score of 80.96, macro F1 score of 37.94). The largest performance gains over the HiSAN model were for histology, for which a Path-BigBird model increased the micro F1 score by 1.44 points and the macro F1 score by 3.55 points. Overall, the results suggest that a Path-BigBird model with a vocabulary derived from well-curated and deidentified data is the best-performing model. CONCLUSION: The Path-BigBird pathology transformer model improves automated information extraction from pathology reports. Although Path-BigBird outperforms Clinical BigBird and HiSAN, these less computationally expensive models still have utility when resources are constrained.

Adaptive language model training for molecular design.

The vast size of chemical space necessitates computational approaches to automate and accelerate the design of molecular sequences to guide experimental efforts for drug discovery. Genetic algorithms provide a useful framework to incrementally generate molecules by applying mutations to known chemical structures. Recently, masked language models have been applied to automate the mutation process by leveraging large compound libraries to learn commonly occurring chemical sequences (i.e., using tokenization) and predict rearrangements (i.e., using mask prediction). Here, we consider how language models can be adapted to improve molecule generation for different optimization tasks. We use two different generation strategies for comparison, fixed and adaptive. The fixed strategy uses a pre-trained model to generate mutations; the adaptive strategy trains the language model on each new generation of molecules selected for target properties during optimization. Our results show that the adaptive strategy allows the language model to more closely fit the distribution of molecules in the population. Therefore, for enhanced fitness optimization, we suggest the use of the fixed strategy during an initial phase followed by the use of the adaptive strategy. We demonstrate the impact of adaptive training by searching for molecules that optimize both heuristic metrics, drug-likeness and synthesizability, as well as predicted protein binding affinity from a surrogate model. Our results show that the adaptive strategy provides a significant improvement in fitness optimization compared to the fixed pre-trained model, empowering the application of language models to molecular design tasks.

Enhancing Adaptive Physics Refinement Simulations Through the Addition of Realistic Red Blood Cell Counts.

Simulations of cancer cell transport require accurately modeling mm-scale and longer trajectories through a circulatory system containing trillions of deformable red blood cells, whose intercellular interactions require submicron fidelity. Using a hybrid CPU-GPU approach, we extend the advanced physics refinement (APR) method to couple a finely-resolved region of explicitly-modeled red blood cells to a coarsely-resolved bulk fluid domain. We further develop algorithms that: capture the dynamics at the interface of differing viscosities, maintain hematocrit within the cell-filled volume, and move the finely-resolved region and encapsulated cells while tracking an individual cancer cell. Comparison to a fully-resolved fluid-structure interaction model is presented for verification. Finally, we use the advanced APR method to simulate cancer cell transport over a mm-scale distance while maintaining a local region of RBCs, using a fraction of the computational power required to run a fully-resolved model.

High Performance Adaptive Physics Refinement to Enable Large-Scale Tracking of Cancer Cell Trajectory.

The ability to track simulated cancer cells through the circulatory system, important for developing a mechanistic understanding of metastatic spread, pushes the limits of today's supercomputers by requiring the simulation of large fluid volumes at cellular-scale resolution. To overcome this challenge, we introduce a new adaptive physics refinement (APR) method that captures cellular-scale interaction across large domains and leverages a hybrid CPU-GPU approach to maximize performance. Through algorithmic advances that integrate multi-physics and multi-resolution models, we establish a finely resolved window with explicitly modeled cells coupled to a coarsely resolved bulk fluid domain. In this work we present multiple validations of the APR framework by comparing against fully resolved fluid-structure interaction methods and employ techniques, such as latency hiding and maximizing memory bandwidth, to effectively utilize heterogeneous node architectures. Collectively, these computational developments and performance optimizations provide a robust and scalable framework to enable system-level simulations of cancer cell transport.

A data-driven approach to modeling cancer cell mechanics during microcirculatory transport.

In order to understand the effect of cellular level features on the transport of circulating cancer cells in the microcirculation, there has been an increasing reliance on high-resolution in silico models. Accurate simulation of cancer cells flowing with blood cells requires resolving cellular-scale interactions in 3D, which is a significant computational undertaking warranting a cancer cell model that is both computationally efficient yet sufficiently complex to capture relevant behavior. Given that the characteristics of metastatic spread are known to depend on cancer type, it is crucial to account for mechanistic behavior representative of a specific cancer's cells. To address this gap, in the present work we develop and validate a means by which an efficient and popular membrane model-based approach can be used to simulate deformable cancer cells and reproduce experimental data from specific cell lines. Here, cells are modeled using the immersed boundary method (IBM) within a lattice Boltzmann method (LBM) fluid solver, and the finite element method (FEM) is used to model cell membrane resistance to deformation. Through detailed comparisons with experiments, we (i) validate this model to represent cancer cells undergoing large deformation, (ii) outline a systematic approach to parameterize different cell lines to optimally fit experimental data over a range of deformations, and (iii) provide new insight into nucleated vs. non-nucleated cell models and their ability to match experiments. While many works have used the membrane-model based method employed here to model generic cancer cells, no quantitative comparisons with experiments exist in the literature for specific cell lines undergoing large deformation. Here, we describe a phenomenological, data-driven approach that can not only yield good agreement for large deformations, but explicitly detail how it can be used to represent different cancer cell lines. This model is readily incorporated into cell-resolved hemodynamic transport simulations, and thus offers significant potential to complement experiments towards providing new insights into various aspects of cancer progression.

Non-invasive characterization of complex coronary lesions.

Conventional invasive diagnostic imaging techniques do not adequately resolve complex Type B and C coronary lesions, which present unique challenges, require personalized treatment and result in worsened patient outcomes. These lesions are often excluded from large-scale non-invasive clinical trials and there does not exist a validated approach to characterize hemodynamic quantities and guide percutaneous intervention for such lesions. This work identifies key biomarkers that differentiate complex Type B and C lesions from simple Type A lesions by introducing and validating a coronary angiography-based computational fluid dynamic (CFD-CA) framework for intracoronary assessment in complex lesions at ultrahigh resolution. Among 14 patients selected in this study, 7 patients with Type B and C lesions were included in the complex lesion group including ostial, bifurcation, serial lesions and lesion where flow was supplied by collateral bed. Simple lesion group included 7 patients with lesions that were discrete, [Formula: see text] long and readily accessible. Intracoronary assessment was performed using CFD-CA framework and validated by comparing to clinically measured pressure-based index, such as FFR. Local pressure, endothelial shear stress (ESS) and velocity profiles were derived for all patients. We validates the accuracy of our CFD-CA framework and report excellent agreement with invasive measurements ([Formula: see text]). Ultra-high resolution achieved by the model enable physiological assessment in complex lesions and quantify hemodynamic metrics in all vessels up to 1mm in diameter. Importantly, we demonstrate that in contrast to traditional pressure-based metrics, there is a significant difference in the intracoronary hemodynamic forces, such as ESS, in complex lesions compared to simple lesions at both resting and hyperemic physiological states [n = 14, [Formula: see text]]. Higher ESS was observed in the complex lesion group ([Formula: see text] Pa) than in simple lesion group ([Formula: see text] Pa). Complex coronary lesions have higher ESS compared to simple lesions, such differential hemodynamic evaluation can provide much the needed insight into the increase in adverse outcomes for such patients and has incremental prognostic value over traditional pressure-based indices, such as FFR.

Limitations of Transformers on Clinical Text Classification.

Bidirectional Encoder Representations from Transformers (BERT) and BERT-based approaches are the current state-of-the-art in many natural language processing (NLP) tasks; however, their application to document classification on long clinical texts is limited. In this work, we introduce four methods to scale BERT, which by default can only handle input sequences up to approximately 400 words long, to perform document classification on clinical texts several thousand words long. We compare these methods against two much simpler architectures - a word-level convolutional neural network and a hierarchical self-attention network - and show that BERT often cannot beat these simpler baselines when classifying MIMIC-III discharge summaries and SEER cancer pathology reports. In our analysis, we show that two key components of BERT - pretraining and WordPiece tokenization - may actually be inhibiting BERT's performance on clinical text classification tasks where the input document is several thousand words long and where correctly identifying labels may depend more on identifying a few key words or phrases rather than understanding the contextual meaning of sequences of text.

Localization of Rolling and Firm-Adhesive Interactions Between Circulating Tumor Cells and the Microvasculature Wall.

INTRODUCTION: The adhesion of tumor cells to vessel wall is a critical stage in cancer metastasis. Firm adhesion of cancer cells is usually followed by their extravasation through the endothelium. Despite previous studies identifying the influential parameters in the adhesive behavior of the cancer cell to a planer substrate, less is known about the interactions between the cancer cell and microvasculature wall and whether these interactions exhibit organ specificity. The objective of our study is to characterize sizes of microvasculature where a deformable circulating cell (DCC) would firmly adhere or roll over the wall, as well as to identify parameters that facilitate such firm adherence and underlying mechanisms driving adhesive interactions. METHODS: A three-dimensional model of DCCs is applied to simulate the fluid-structure interaction between the DCC and surrounding fluid. A dynamic adhesion model, where an adhesion molecule is modeled as a spring, is employed to represent the stochastic receptor-ligand interactions using kinetic rate expressions. RESULTS: Our results reveal that both the cell deformability and low shear rate of flow promote the firm adhesion of DCC in small vessels ( < 10 µ m ). Our findings suggest that ligand-receptor bonds of PSGL-1-P-selectin may lead to firm adherence of DCC in smaller vessels and rolling-adhesion of DCC in larger ones where cell velocity drops to facilitate the activation of integrin-ICAM-1 bonds. CONCLUSIONS: Our study provides a framework to predict accurately where different DCC-types are likely to adhere firmly in microvasculature and to establish the criteria predisposing cancer cells to such firm adhesion.

Multi-GPU Immersed Boundary Method Hemodynamics Simulations.

Large-scale simulations of blood flow that resolve the 3D deformation of each comprising cell are increasingly popular owing to algorithmic developments in conjunction with advances in compute capability. Among different approaches for modeling cell-resolved hemodynamics, fluid structure interaction (FSI) algorithms based on the immersed boundary method are frequently employed for coupling separate solvers for the background fluid and the cells within one framework. GPUs can accelerate these simulations; however, both current pre-exascale and future exascale CPU-GPU heterogeneous systems face communication challenges critical to performance and scalability. We describe, to our knowledge, the largest distributed GPU-accelerated FSI simulations of high hematocrit cell-resolved flows with over 17 million red blood cells. We compare scaling on a fat node system with six GPUs per node and on a system with a single GPU per node. Through comparison between the CPU- and GPU-based implementations, we identify the costs of data movement in multiscale multi-grid FSI simulations on heterogeneous systems and show it to be the greatest performance bottleneck on the GPU.

Determining the impacts of venoarterial extracorporeal membrane oxygenation on cerebral oxygenation using a one-dimensional blood flow simulator.

Venoarterial extracorporeal membrane oxygenation (VA-ECMO) is a mechanical system that provides rapid and short-term support for patients with cardiac failure. In many patients, pulmonary function is also impaired, resulting in poorly-oxygenated cardiac outflow competing against well-oxygenated VA-ECMO outflow, a condition known as North-South syndrome. North-South syndrome is a primary concern because of its potential to cause cerebral hypoxia, which has a critical influence on neurological complications often seen in this patient population. In order to reduce ischemic neurological complications, it is important to understand how clinical decisions regarding VA-ECMO parameters influence blood oxygenation. Here, we studied the impacts of flow rate and cannulation site on oxygenation using a one-dimensional (1D) model to simulate blood flow. Our model was initially tested by comparing blood flow results to those observed from experimental work in VA-ECMO patients. The 1D model was combined with a two-phase flow model to simulate oxygenation. Additionally, the influence of various other clinician-tunable parameters on oxygenation in the common carotid arteries (CCAs) were tested, including, blood viscosity, cannula position within the insertion artery, heart rate, and systemic vascular resistance (SVR), as well as geometrical changes such as arterial radius and length. Our results indicated that blood oxygenation to the brain strongly depended on the cannula insertion site and the VA-ECMO flow rate with a weaker but potentially significant dependence on arterial radius. During femoral cannulation, VA-ECMO flow rates greater than ~4.9L/min were needed to perfuse the CCAs. However, axillary and central cannulation began to perfuse the CCAs at significantly lower flow (~1L/min). These results may help explain the incidence of cerebral hypoxia in this patient population and the common need to change cannulation strategies during treatment to address this clinical problem. While this work describes patient-averaged results, determining these relationships between VA-ECMO parameters and cerebral hypoxia is an important step towards future work to develop patient-specific models that clinicians can use to improve outcomes.

Accelerating massively parallel hemodynamic models of coarctation of the aorta using neural networks.

Comorbidities such as anemia or hypertension and physiological factors related to exertion can influence a patient's hemodynamics and increase the severity of many cardiovascular diseases. Observing and quantifying associations between these factors and hemodynamics can be difficult due to the multitude of co-existing conditions and blood flow parameters in real patient data. Machine learning-driven, physics-based simulations provide a means to understand how potentially correlated conditions may affect a particular patient. Here, we use a combination of machine learning and massively parallel computing to predict the effects of physiological factors on hemodynamics in patients with coarctation of the aorta. We first validated blood flow simulations against in vitro measurements in 3D-printed phantoms representing the patient's vasculature. We then investigated the effects of varying the degree of stenosis, blood flow rate, and viscosity on two diagnostic metrics - pressure gradient across the stenosis (ΔP) and wall shear stress (WSS) - by performing the largest simulation study to date of coarctation of the aorta (over 70 million compute hours). Using machine learning models trained on data from the simulations and validated on two independent datasets, we developed a framework to identify the minimal training set required to build a predictive model on a per-patient basis. We then used this model to accurately predict ΔP (mean absolute error within 1.18 mmHg) and WSS (mean absolute error within 0.99 Pa) for patients with this disease.

Suitability of lattice Boltzmann inlet and outlet boundary conditions for simulating flow in image-derived vasculature.

The lattice Boltzmann method (LBM) is a popular alternative to solving the Navier-Stokes equations for modeling blood flow. When simulating flow using the LBM, several choices for inlet and outlet boundary conditions exist. While boundary conditions in the LBM have been evaluated in idealized geometries, there have been no extensive comparisons in image-derived vasculature, where the geometries are highly complex. In this study, the Zou-He (ZH) and finite difference (FD) boundary conditions were evaluated in image-derived vascular geometries by comparing their stability, accuracy, and run times. The boundary conditions were compared in four arteries: a coarctation of the aorta, dissected aorta, femoral artery, and left coronary artery. The FD boundary condition was more stable than ZH in all four geometries. In general, simulations using the ZH and FD method showed similar convergence rates within each geometry. However, the ZH method proved to be slightly more accurate compared with experimental flow using three-dimensional printed vasculature. The total run times necessary for simulations using the ZH boundary condition were significantly higher as the ZH method required a larger relaxation time, grid resolution, and number of time steps for a simulation representing the same physiological time. Finally, a new inlet velocity profile algorithm is presented for complex inlet geometries. Overall, results indicated that the FD method should generally be used for large-scale blood flow simulations in image-derived vasculature geometries. This study can serve as a guide to researchers interested in using the LBM to simulate blood flow.

The importance of side branches in modeling 3D hemodynamics from angiograms for patients with coronary artery disease.

Genesis of atherosclerotic lesions in the human arterial system is critically influenced by the fluid mechanics. Applying computational fluid dynamic tools based on accurate coronary physiology derived from conventional biplane angiogram data may be useful in guiding percutaneous coronary interventions. The primary objective of this study is to build and validate a computational framework for accurate personalized 3-dimensional hemodynamic simulation across the complete coronary arterial tree and demonstrate the influence of side branches on coronary hemodynamics by comparing shear stress between coronary models with and without these included. The proposed novel computational framework based on biplane angiography enables significant arterial circulation analysis. This study shows that models that take into account flow through all side branches are required for precise computation of shear stress and pressure gradient whereas models that have only a subset of side branches are inadequate for biomechanical studies as they may overestimate volumetric outflow and shear stress. This study extends the ongoing computational efforts and demonstrates that models based on accurate coronary physiology can improve overall fidelity of biomechanical studies to compute hemodynamic risk-factors.