Large Scale Lattice Boltzmann for Biocolloidal Systems

Lead Research Organisation: University College London
Department Name: Chemistry

Abstract

Many of the most important structural components of biological materials exist on length scales of nanometres to microns. Examples include long polymers (such as DNA or flexible proteins), compact objects such as globular proteins (either in isolation or in structured assemblies such as the body of a virus); and lipid bilayer membranes (such as the walls that enclose the cells in our bodies). This range of length-scales is known as the colloidal domain. It is one where physical processes can be at least as important as biochemical ones; for instance, inside every cell, there are 'regulatory' proteins which have to search the genome (DNA) seeking preferential places of attachment (from which they then control the production of other proteins). To find these places, the regulatory proteins depend, at least in part, on the completely random motion, called Brownian motion, that occurs when particles on the colloidal scale are bombarded by the thermal energy of the surrounding solvent (water) molecules. This diffusive process is considerably complicated by the fact that motion of one colloidal object sets the surrounding solvent into motion, which causes all nearby objects also to move. This is referred to as a hydrodynamic interaction. Similarly, if one designs a drug delivery system in which drug molecules are encapsulated, the capsules are again on the colloidal scale but their motion in the bloodstream is dominated by their being swept along by the flow of blood, which is another form of hydrodynamic effect. In some modern therapies, magnetic colloids are steered with or against this flow by an external field; in such cases it is important to understand the effect of hydrodynamics on the response to a force.The physical consequences of hydrodynamic couplings in bio-colloidal systems are thus wide ranging. However, it is currently very difficult to predict any of these important effects, even using simplified models in which the biochemical detail of the colloidal objects is omitted. Fortunately, such problems can increasingly be addressed using very large scale computer simulation on some of the world's most powerful computers. The so-called lattice Boltzmann algorithm (LB) offers a specific technical solution to the challenges of hydrodynamics, by using a discrete lattice to model the flow of fluid from place to place. Unlike some other methods it can include the random forces responsible for Brownian motion. Also, as well as modelling colloids and polymers surrounded by simple solvents such as water, LB can also address solvents comprising complex fluids. The latter include the so-called 'amphiphilic mesophases' in which small molecules (with a water-loving head and water-hating tail) self-assemble into a labyrinth within which proteins or nanocolloids can reside. We aim to develop LB algorithms in the bio-colloidal context, and apply these to create new scientific knowledge that has been out of reach until now. Indeed we plan very large simulations of a range of hydrodynamic problems that lie at the interface between physics and the life sciences. These problems include: the flow of magnetic colloids in the blood stream as a model for 'Magnetic Drug Targetting' (MDT) on real patients; the motion of nanocolloids within amphiphilic mesophases suitable for drug delivery applications; the ejection of DNA from the body of a virus as it infects a cell; the dynamical behaviour of the highly confined DNA that is found within the bacterial and other cells; and the interaction between colloidal particles and DNA within the cellular environment. The last of these topics is crucial to understanding the problem of genome exploration by regulatory proteins as mentioned earlier, which we also plan to address. In all these areas, large-scale computer simulation has the potential to change the way science is done. We hope to establish a world lead for the UK in this emerging field.

Planned Impact

The research planned in this proposal addresses areas where large, accurate simulation models can shed crucial light on the physical processes determining key aspects of biological function. Our adventurous scientific programme will have widespread benefits in pure and applied science, engineering, industry and health. Our proposed simulations on vascular flow and magnetic drug targetting (MDT) build directly on Coveney's GENIUS HemeLB project which involves clinical collaborators and the handling of patient-specific data on the layout of blood vessels. Development of patient-specific vascular simulation for MDT, if it can become part of established clinical practice, will be of obvious direct benefit to patients; it will also empower the clinicians who treat them, and potentially save costs by increasing success rates in complex treatment procedures. The work on colloidal transport in amphiphilic labyrinths will inform a wider body of ongoing research aimed at developing encapsulation technologies. Here the potential to benefit patients and their clinicians also exists in the longer term. The delivery systems modelled in this project may also have other commercially important applications, for instance in food technology and personal care products where controlled release of an encapsulated active agent is required. Our large scale simulations of DNA and biocolloid hydrodynamics in realistic geometries, including the interior of living cells, will be relevant to the many researchers now working at the physics / life sciences interface, including cell and bacterial biologists, biomedical engineers and possibly some clinicians. Although the majority of such researchers are academic, there are also quite large numbers in the commercial sector (particularly pharmaceuticals but increasingly also foodstuffs, remediation and biotech sectors). Alongside the biocolloidal applications that are the focus of this project, the lattice-Boltzmann simulation capability that we plan to develop could benefit a range of industrial sectors. For instance, recent Edinburgh codes, developed to address the design of new colloid-stabilized emulsion materials, were subsequently redeployed by researchers at Schlumberger to model multicomponent flows in porous media. Likewise UCL has ongoing collaborations, not only with the National Hospital for Neurology and Neurosurgery in haemodynamics simulation, but also with MI-SWACO in the area of optimising properties of drilling fluids for oil recovery. Thus our planned work on further developing lattice Boltzmann methods for colloids and polymers, as well as impacting scientifically on a wide range of stakeholders at the physics / life science interface through the specific scientific projects detailed in this proposal, has wider potential impact through development of new computational capabilities and software infrastructure. The latter is relevant to a broad class of colloidal and complex fluid materials used in industries such as chemicals, foodstuffs, personal care, and oil recovery.

Publications

10 25 50

publication icon
Baxter R (2013) EUDAT: A New Cross-Disciplinary Data Infrastructure for Science in International Journal of Digital Curation

publication icon
Bernabeu MO (2015) Characterization of parafoveal hemodynamics associated with diabetic retinopathy with adaptive optics scanning laser ophthalmoscopy and computational fluid dynamics. in Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference

publication icon
Bhati AP (2018) Uncertainty Quantification in Alchemical Free Energy Methods. in Journal of chemical theory and computation

publication icon
Borgdorff J (2014) Performance of distributed multiscale simulations. in Philosophical transactions. Series A, Mathematical, physical, and engineering sciences

 
Description The project has enabled the development and application of two kinetic lattice-Boltzmann (fluid particle dynamics solvers) models for the simulation of colloidal particles in biological systems on different scales.
The program HemeLB, optimised for sparse vascular geometries, has been extended to include magnetic colloidal particles. This was achieved by combining the fluid solver with point particle model coupled by a fluid friction model.
HemeLB has been evaluated and applied to various biomedical problems involving blood flow (hemodynamics). An evaluation of wall boundary conditions for the modelling of blood flow has elucidated the applicability of the method for different flow situations as well as resolution requirements for stability and accuracy.
Subsequently the model was employed for the investigation of the generation of vascular (blood vessel) networks (angiogenesis). Simulations of the flow field in developing mouse retina were directly compared to experimental findings. Numerical investigations helped to elucidate the interplay of wall shear stress and axial polarization of endothelial cells (cells of the blood vessel wall) in developing vascular networks.
Utilising the program LB3D established for the large-Scale simulation of amphiphilic fluid mixtures (containing three phases including one surface active phase, e.g. water, oil and soap) properties of a coupled particle models have been investigated. Interactions of explicit spherical and ellipsoidal particles with binary interfaces and external fields have been studied with focus on fluid interface self-organisation and dynamics.
In conjunction with computing resources provided by the EPSRC UK high end computing consortium on mesoscale engineering sciences (UKCOMES, EP/L00030X) both work on angiogenesis as well as complex fluid interface dynamics has produced high impact publications in internationally acclaimed journals.
Exploitation Route The developed mesoscopic models for the simulation of colloidal particles under the influence of external force fields in blood flow and other complex fluids allow the simulation of a wealth of systems relevant to biomedical as well as industrial applications involving complex fluid mixtures (E.g. cosmetics, food and oil production).
Consequent continuation of development and application of both the sparse network model of HemeLB and the explicit particle model in LB3D is highly relevant to the investigation of drug targeting techniques using magnetic particles and fields as well as in the development of novel drug carriers based on self-organisation phenomena.
Sectors Digital/Communication/Information Technologies (including Software),Manufacturing, including Industrial Biotechology,Pharmaceuticals and Medical Biotechnology

 
Description This project has facilitated the inclusion of LB3D in the petascale application library of Fujitsu UK. Fujistsu UK funded 50% of the Ph.D. student Gary Davies working in the EPSRC Biocolloids project (EP/I034602) and high end computing consortium UKCOMES (EP/L00030X). The project furthermore informed decisions of Prof Coveney both in context of his chairing of the CPP steering panel and his work on the Strategy for the UK Research Computing Ecosystem. The latter lead to the establishment of the UK e-infrastructure leadership council which has been influential concerning a UK government investment volume of £500M into HPC resources over the period 2011-2016.
Sector Digital/Communication/Information Technologies (including Software),Government, Democracy and Justice
Impact Types Economic,Policy & public services

 
Description FETHPC-1-2014 - HPC Core Technologies, Programming Environments and Algorithms for Extreme Parallelism and Extreme Data Applications
Amount € 4,122,864 (EUR)
Funding ID 671564 
Organisation European Commission H2020 
Sector Public
Country Belgium
Start 10/2015 
End 09/2018
 
Title FabSim 
Description We present FabSim, a toolkit developed to simplify a range of computational tasks for researchers in diverse disciplines. FabSim is flexible, adaptable, and allows users to perform a wide range of tasks with ease. It also provides a systematic way to automate the use of resourcess, including HPC and distributed resources, and to make tasks easier to repeat by recording contextual information. To demonstrate this, we present three use cases where FabSim has enhanced our research productivity. These include simulating cerebrovascular bloodflow, modelling clay-polymer nanocomposites across multiple scales, and calculating ligand-protein binding affinities. 
Type Of Technology Software 
Year Produced 2016 
Open Source License? Yes  
Impact FabSim has been used in a range of journal publications in bloodflow modelling, clay-polymer modelling, and modelling of protein-ligand binding affinities. 
URL http://www.github.com/UCL-CCS/FabSim
 
Title HemeLB 
Description The code is an open-source, parallel, lattice-Boltzmann blood flow simulator developed at the Centre for Computational Science (CCS) at UCL. It is one of the three main pieces of open source software of UKCOMES. 
Type Of Technology Software 
Year Produced 2015 
Open Source License? Yes  
Impact The HemeLB suite is able to generate 3D models of the vasculature of individual human body parts based on medical images such as an angiogram, MRI or CT scan. These models are then used to run sophisticated fluid dynamics simulations (using the Lattice Boltzmann method) which can provide accurate haemodynamic estimates for blood vessels; for example, blood pressure, flow rate, and wall shear stress at different locations. 
URL https://www.ucl.ac.uk/research-it-services/case-studies-pub/rsdt-case-study-4
 
Title LB3D 
Description LB3D is an open-source code for simulating three-dimensional simple, binary oil/water and ternary oil/water/amphiphile fluids using the Shan-Chen model for binary fluid interactions. It is written in Fortran 90 and parallelized using MPI. It supports XDR and HDF5 format for I/O. It is one of the three main pieces of open source software of UKCOMES. 
Type Of Technology Software 
Year Produced 2016 
Open Source License? Yes  
Impact It has been used by a number of research groups in the world. LB3D has been used to study self-assembly of cubic phases, micro-mixing, flow through porous media, fluid surface interactions and other problems in complex fluidics. 
URL https://ccpforge.cse.rl.ac.uk/gf/project/lb3d
 
Title LB3D: Lattice-Boltzmann three dimensional simulation of fluids 
Description LB3D provides functionality to simulate three-dimensional simple, binary oil, water and ternary oil, water and amphiphile fluids using the Shan-Chen model for binary fluid interactions. The boundary conditions available include periodic boundaries, body forcing, and bounce-back boundaries as Lees-Edwards shearing for simple and binary fluid mixtures. The software is written in Fortran 90 and parallelized using MPI. It supports XDR and HDF5 format for I/O and provides checkpoint and restart for long-running simulations. It has been ported to many supercomputers worldwide, where it has shown excellent scalability. Most recently it has been shown to scale linearly on up to 294,000 cores on the European Blue Gene/P system Jugene. LB3D is developed by the Centre for Computational Chemistry at University College London, University of Stuttgart, and the Technical University of Eindhoven. EPCC has contributed to the development of LB3D most recently as part of the "Biocolloids" project funded by EPSRC. 
Type Of Technology Software 
Year Produced 2015 
Open Source License? Yes  
Impact LB3D has been used to study self-assembly of cubic phases, micro-mixing, flow through porous media, fluid surface interactions and other problems in complex fluidics. In addition, we have implemented the moment propagation method in LB3D, which allows us to efficiently measure the effective diffusion, dispersion, and other transport properties of ionic species in electrokinetic flows. These developments have enabled us to reveal interesting phenomena in charged rocks, namely a non-monotonic dispersion coefficient and a crossover in the effective diffusion of charged tracers. The publication of these novel scientific insights is currently under preparation [Schiller and Coveney]. This research has benefitted considerably from the generous allocation of computing time on ARCHER through UKCOMES. Dr. Schiller will continue to develop and use LB3D at Clemson University, USA, thus broadening the international scientific impact of LB3D. Furthermore, the LB3D code is one of the application codes in the European H2020 Compat project http://www.compat-project.eu/, where it will be used to enable bleeding edge simulations for computational materials science research on emerging exascale HPC systems. 
URL http://ccpforge.cse.rl.ac.uk/gf/project/lb3d
 
Description Press release, national news articles 
Form Of Engagement Activity A press release, press conference or response to a media enquiry/interview
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Public/other audiences
Results and Impact As a result of our research and subsequent academic publication in Advanced Materials, University College London released a press release that was subsequently picked up by various news outlets: A Telegraph article entitled "New 'Virtual Laboratory' will change how we approach material chemistry", concerned this research on predicting the large-scale properties of materials based on their chemical composition. A BBC website article, "Clay composites modelled in 'virtual lab'", was based on the work published in our Advanced Materials feature article. This research was also highlighted on the popular science/technology website The Register.

This work also featured in the PRACE (Partnership for Advanced Computing in Europe) newsletter and won first prize in the 2015 Journal of Polymer Science award in the Theory and Modeling of Nanoparticles: Interactions with Biomolecules and Polymers Session at the American Chemical Society Meeting, Fall 2015.
Year(s) Of Engagement Activity 2014