No abstract

BackgroundVirtual patients are increasingly common tools used in health care education to foster learning of clinical reasoning skills. One potential way to expand their functionality is to augment virtual patients’ interactivity by enriching them with computational models of physiological and pathological processes.ObjectiveThe primary goal of this paper was to propose a conceptual framework for the integration of computational models within virtual patients, with particular focus on (1) characteristics to be addressed while preparing the integration, (2) the extent of the integration, (3) strategies to achieve integration, and (4) methods for evaluating the feasibility of integration. An additional goal was to pilot the first investigation of changing framework variables on altering perceptions of integration.MethodsThe framework was constructed using an iterative process informed by Soft System Methodology. The Virtual Physiological Human (VPH) initiative has been used as a source of new computational models. The technical challenges associated with development of virtual patients enhanced by computational models are discussed from the perspectives of a number of different stakeholders. Concrete design and evaluation steps are discussed in the context of an exemplar virtual patient employing the results of the VPH ARCH project, as well as improvements for future iterations.ResultsThe proposed framework consists of four main elements. The first element is a list of feasibility features characterizing the integration process from three perspectives: the computational modelling researcher, the health care educationalist, and the virtual patient system developer. The second element included three integration levels: basic, where a single set of simulation outcomes is generated for specific nodes in the activity graph; intermediate, involving pre-generation of simulation datasets over a range of input parameters; advanced, including dynamic solution of the model. The third element is the description of four integration strategies, and the last element consisted of evaluation profiles specifying the relevant feasibility features and acceptance thresholds for specific purposes. The group of experts who evaluated the virtual patient exemplar found higher integration more interesting, but at the same time they were more concerned with the validity of the result. The observed differences were not statistically significant.ConclusionsThis paper outlines a framework for the integration of computational models into virtual patients. The opportunities and challenges of model exploitation are discussed from a number of user perspectives, considering different levels of model integration. The long-term aim for future research is to isolate the most crucial factors in the framework and to determine their influence on the integration outcome.

BackgroundThis study combined themes in cardiovascular modelling, clinical cardiology and e-learning to create an on-line environment that would assist undergraduate medical students in understanding key physiological and pathophysiological processes in the cardiovascular system.MethodsAn interactive on-line environment was developed incorporating a lumped-parameter mathematical model of the human cardiovascular system. The model outputs were used to characterise the progression of key disease processes and allowed students to classify disease severity with the aim of improving their understanding of abnormal physiology in a clinical context. Access to the on-line environment was offered to students at all stages of undergraduate training as an adjunct to routine lectures and tutorials in cardiac pathophysiology. Student feedback was collected on this novel on-line material in the course of routine audits of teaching delivery.ResultsMedical students, irrespective of their stage of undergraduate training, reported that they found the models and the environment interesting and a positive experience. After exposure to the environment, there was a statistically significant improvement in student performance on a series of 6 questions based on cardiovascular medicine, with a 33% and 22% increase in the number of questions answered correctly, p < 0.0001 and p < 0.001 respectively.ConclusionsConsiderable improvement was found in students’ knowledge and understanding during assessment after exposure to the e-learning environment. Opportunities exist for development of similar environments in other fields of medicine, refinement of the existing environment and further engagement with student cohorts. This work combines some exciting and developing fields in medical education, but routine adoption of these types of tool will be possible only with the engagement of all stake-holders, from educationalists, clinicians, modellers to, most importantly, medical students.Electronic supplementary materialThe online version of this article (10.1186/s12909-017-1058-1) contains supplementary material, which is available to authorized users.

A crucial enabling technology for structural genomics is the development of algorithms that can predict the putative function of novel protein structures: the proposed functions can subsequently be experimentally tested by functional studies. Testable assignments of function can be made if it is possible to attribute a putative, or indeed probable, function on the basis of the shapes of the binding sites on the surface of a protein structure. However the comparison of the surfaces of 3D protein structures is a computationally demanding task. Here we present four surface representations that can be used locally to describe the global shape of specifically bounded local region models. The most successful of these representations is obtained by a Fourier analysis of the distribution of surface curvature on concentric spheres around a surface point and summarizes a 24 A diameter spherically clipped region of protein surface by a fingerprint of 18 Fourier amplitude values. Searching experiments using these fingerprints on a set of 366 proteins demonstrate that this provides an effective and an efficient technique for the matching of protein surfaces.

We report the development and validation of the program GENFOLD, a genetic algorithm that calculates protein structures using restraints obtained from NMR, such as distances derived from nuclear Overhauser effects, and dihedral angles derived from coupling constants. The program has been tested on three proteins: the POU domain (a small three-helix DNA-binding protein), bovine pancreatic trypsin inhibitor (BPTI), and the starch-binding domain from Aspergillus niger glucoamylase I, a 108-residue P-sheet protein. Structures were calculated for each protein using published NMR restraints. In addition, structures were calculated for BPTI using artificial restraints generated from a high-resolution crystal structure. In all cases the fittest calculated structures were close to the target structure, and could be refined to structures indistinguishable from the target structures by means of a low-temperature simulated annealing refinement. The effectiveness of the program is similar to that of distance geometry and simulated annealing methods, and it is capable of using a very wide range of restraints as input. It can thus be readily extended to the calculation of structures of large proteins, for which few NOE restraints may be available.Keywords: genetic algorithm; NMR; protein folding; protein structure; simulated annealing NMR is a widely used technique for obtaining information on protein structures in solution (Wuthrich, 1986). The information usually takes the form of distances, specified as distance ranges derived from nuclear Overhauser effect (NOE) information (Neuhaus & Williamson, 1989), and angles, also specified as ranges, and obtained from spin-spin coupling constants and from "C chemical shifts. In addition, hydrogen bonds can be specified based on previous assignments of locations of regular secondary structure elements, and can be applied as distance restraints. These measures form the NMR restraints on the structure, which are supplemented by a much larger number of "holonomic" constraints, consisting of bond lengths and angles, atomic radii, etc. Several methods are used for applying the restraints to generate structures, of which the most common methods are simulated annealing and distance geometry (which is often applied as a front end to simulated annealing: Nilges et al., 1988). The term "distance geometry" is applied to two quite different methods: the metric matrix method, such as the program DGII, which works initially in distance space and subsequently embeds the structure into Cartesian space (Have1 & Reprint requests to: M.P.