Please see below the basic format of the programme for each day. You will see that some symposia have been amalgamated and some are over more than one day.
Most musculoskeletal system models appeal to multi-body simulation frameworks, in which the skeletal muscle force generation is modelled using Hill-type skeletal muscle models. Such modelling frameworks have the advantage that they can analyse and predict movement using musculoskeletal system models with a realistic number of muscle (groups). However, these multi-body simulation frameworks are also based on limiting modelling assumptions. For example, Hill-type skeletal muscle models lump together anatomical and physiological complexity to a few lumped parameters, e.g., the complex muscle fiber architecture to a single parameter at one point in space. Therefore, Hill-type models cannot be used to investigate key phenomena such as, for example, contact with external objects or muscle-muscle or muscle-bone interaction. Musculoskeletal system models appealing to three-dimensional, continuum-mechanical skeletal muscle models could naturally overcome such limitations. However, such models are rare and require sophisticated constitutive models and large computational resources. This is particularly true for forward (dynamics) simulations that are based on solving optimization problems. Based on the currently only published forward simulations of a two-muscle upper limb model, which consists of a Biceps Brachii, Triceps Brachii, the Humerus, and a one-degree-of-freedom elbow joint, cf. [1, 2] and Figure 1 (left), we will discuss the particular challenges in modelling musculoskeletal system models consisting of multiple continuum-mechanical, three-dimensional skeletal muscle models (cf. Figure 1 (right)).Full Abstract
Artificial intelligence (AI) is emerging as an important new approach to how new medicines will be designed. The first molecules designed by algorithm are heading for the clinic and we now have evidence on the impact AI technologies will have on the drug discovery process. Originally, spun out from his lab at the University of Dundee (Nature 492(7428):215-20),
Exscientia is the first company to demonstrate the automation of drug design surpassing human endeavour. Whilst a wealth of machine-readable data in chemistry, pharmacology and biology is now available, drug discovery, especially for novel drug targets, where little or no pharmacology data exists, is by definition a “small data” problem. Moreover, the problem of data annotation (via experiments) is relatively slow and expensive. Slow and expensive. Here we will introduce how small amounts of seed data can be generated and how the development of active learning algorithms to learn efficiently from small data is the appropriate strategy to apply AI to the design of first in class drugs. The power of the approach will be demonstrated in cases studies of the pre-clinical drug candidates that have been designed by the system that are now in late stage IND-enabling studies. Exscientia’s productivity metrics achieved on its first projects indicate the application of AI to drug discovery can potentially reducing the cost of bring a drug to market by 30%, approximately around $600m per licensed drug.Full Abstract
Cerebral aneurysms are abnormal enlargements of the walls of brain arteries that are typically saccular in shape and found largely at arterial bifurcations in the vicinity of an anatomic structure at the base of the brain called the Circle of Willis. Rupture of a cerebral aneurysm is a central cause of subarachnoid hemorrhage, a devastating type of stroke with high mortality and disability rates. As the majority of aneurysms do not rupture during a person’s lifetime and treatments for unruptured aneurysms have serious medical risks, there is an urgent need to develop reliable methods for assessing the likelihood of rupture.Full Abstract
Drug discovery is being pursued through computer-aided design, synthesis, biological assaying, and crystallography.1-4 Lead identification features de novo design with the ligand growing program BOMB or virtual screening. Emphasis is placed on optimization of the resultant hits to yield potent, drug-like inhibitors. Monte Carlo/free-energy perturbation (FEP) simulations are often executed to identify the most promising choices for substituents on rings, heterocycles, and linking groups. The illustrated applications center on the design of inhibitors targeting HIV-1 reverse transcriptase, macrophage migration inhibitory factor, and JAK2 kinase. Micromolar leads have been rapidly advanced to low nanomolar inhibitors, and numerous crystal structures for protein-inhibitor complexes have been obtained. Development and use of fluorescence polarization assays provide direct binding data. Key computational issues are considered including force fields, atomic charge models, conformational sampling, computation of absolute free energies of binding, and unbinding pathways from metadynamics.Full Abstract
16:15 – 16:30
Introducing Sano – Centre for Computational Personalised Medicine – International Research Foundation,
Andrew Narracott and Piotr Nowakowski
A new scientific entity called Sano – Centre for Computational Personalised Medicine – will be established in Kraków, Poland. This international research foundation is one of three Polish beneficiaries of the Teaming for Excellence. The mission of Sano involves:
development of new computational methods, algorithms, models and technologies for personalized medicine,
introducing new diagnostic and therapeutic solutions based on computerized simulations into clinical practice,
fostering creation and growth of enterprises which develop cutting-edge diagnostic and therapeutic technologies,
contributing to novel training and education curricula which meet the needs of modern personalised medicine.<a