(6am) Multi-Scale Biomolecular Modeling and Design for Engineering and Medicine

Research Interests:

My research focuses on multi-scale modeling of proteins and leveraging these models to gain insights into disease, as well as to design potential therapeutics. Biophysical and physicochemical understanding of protein function is a central element of my work, but one theme of my research has been developing models across physiological scales by using data mining and machine learning to integrate multiple types of experimental and computational data. A primary application area of my work has been computational immunology, where I have been involved in projects focused on topics such as the complement immune system and HIV. My recent research has focused on a family of proteases known to be upregulated in various types of cancer, but the development of tools for general protein structure prediction and protein design is also of great interest. My recent work, has started to combine molecular modeling with systems-level models to better understand the contribution of molecular changes in complex protein networks. One major interest of mine is the development of computational tools for research and teaching, including the development of multiple R packages and more recently the development of intuitive online interfaces to allow nonexperts to interact with models using the R Shiny platform.

Research Experience:

The core of my training and experience has been in computational modeling, starting with an emphasis biophysical models of proteins when I was a graduate student, including continuum electrostatics and molecular dynamics simulations. My graduate work also included computational drug design based on computational chemistry methods, such as pharmacophore searches and virtual screening. As a postdoctoral researcher, my work remained computational, but transitioned to an approach more centered on data science, utilizing machine learning and global optimization. My current project still involves biophysical modeling for understanding protein interactions, but also includes a systems modeling component for which I have also been developing an algorithm for parameter estimation. In addition to having the opportunity work on a diverse collection of computational projects, throughout my academic career I have also had an interest in coupling computational modeling with experiments. This interest began as a graduate student where I used NMR spectroscopy to characterize peptide structure, as well as, gaining experience in measuring the potency of designed inhibitors using ELISAs. As a postdoctoral researcher, one of my primary projects involved performing solid-phase peptide synthesis for over 40 peptides, that where then used in high-throughput peptide solubility measurements. More recently I have gained experience with basic cell and molecular biology techniques including gel electrophoresis, Western blots, and tissue culture.

Successful Proposals: UCOP Tobacco Related Disease Research Program Dissertation Fellowship, NSF/DAAD Central Europe Summer Research Institute Research Fellowship, NSF/JSPS East Asia and the Pacific Summer Institute Research Fellowship.

Current Project: “Computational modeling of cysteine cathepsin proteases and the proteolytic network.”

Under the supervision of Manu O. Platt, Coulter Department of Biomedical Engineering,
Georgia Institute of Technology

Postdoctoral Project: “Application of global optimization and machine learning to predict protein structure and function.”

Under the supervision of Christodoulos A. Floudas, Department of Chemical and Biological Engineering, Princeton University (2012-2014) and Texas A&M Energy Institute (2014-2016)

PhD Dissertation: “Design and Applications of a Computational Framework for Protein and Drug Design.”

Under supervision of Dimitrios Morikis, Department of Bioengineering, University of California, Riverside

Teaching Interests:

For my current appointment, in addition to my research I have also been teaching an undergraduate course for junior and senior biomedical engineering students, Biomedical Systems and Modeling, that introduces the fundamentals of modeling biomedical systems. The course covers a wide variety of topics from system biology, and success in the class requires a grasp of modern biology, analyzing graphs and networks, models consisting of nonlinear differential equations, several types of simulations, and a variety of auxiliary mathematical methods. The course has a pseudo-flipped structure with weekly overview lectures and problem solve sessions that give the students hands-on modeling experience. To complement the problem solving sessions, I have also initiated and been leading an effort to develop new online teaching tools that cover various modeling topics. The students have found these web-based applications very useful in learning difficult concepts in Biomedical Systems and Modeling, since they provide simple interfaces for manipulating models of some key biomedical systems from the course. These teaching tools are not only being used by the students in my sections, but now are also being used by all of the instructors and students in Biomedical Systems and Modeling.

Future Direction:

Future areas of interest are centered around multi-scale biomolecular engineering, and fall into three main areas: 1) data-driven models of biology and disease for personalized medicine; 2) modeling of protein structure and dynamics across atomistic and systems scales; 3) design of peptides/proteins for therapeutic and diagnostic applications. In area 1, I hope to build upon my experience with models for HIV co-receptor selection, and branch out into areas such as predicting antibody-epitope affinities and proteolytic cleavage of peptides. In area 2, early projects will focus on using data-driven methods to build multi-scale models of large protein complexes, such as though found in the complement immune system, and there will also be an emphasis on using biophysical models of protein interactions to predict parameters of system-scale models. Proposed work in area 3, will be both experimental and computational in nature, focusing on the design on multi-functional peptides and initial characterization using basic assays and structural methods. The proposed multi-scale design approach will account for microenvironmental factors, such as pH, and will incorporate functional characteristics such as binding, solubility, and conformational switching, to name a few. Application areas of interest will include computational immunology and proteases in the context of cancer, which will build off of my prior experience. A major component of the proposed work will center around the development of computational tools, and therefore experimental collaborations will also be actively sought in other related areas.

Selected Publications (34 total co-authored publications/reviews):

CA Kieslich, F Boukouvala, CA Floudas (2018) Optimization of black-box problems using Smolyak grids and polynomial approximations. J. Glob. Opt., In press.

CA Kieslich, P Tamamis, YA Guzman, M Onel, CA Floudas (2015) Highly accurate structure-based prediction of HIV-1 coreceptor usage suggests intermolecular interactions driving tropism. PLOS ONE, 11(2), e0148974.

CA Kieslich, J Smadbeck, GA Khoury, CA Floudas (2016) conSSert: Consensus SVM model for accurate prediction of ordered secondary structure, J. Chem. Inf. Mod., 56 (3), 455–46.

CA Kieslich, D Shin, A López de Victoria, G González-Rivera, D Morikis (2013) A predictive model for HIV-1 co-receptor selectivity. AIDS Res. Hum. Retroviruses, 29:1386-1394.

CA Kieslich and D Morikis (2012) The two sides of complement C3d: Evolution of electrostatics in a link between innate and adaptive immunity. PLoS Comp. Bio., 8(12): e1002840.

K Chae, CA Kieslich, D Morikis, SC Kim, EM Lord (2009) A gain-of-function mutation of Arabidopsis lipid transfer protein 5 disturbs pollen tube tip growth and fertilization. Plant Cell, 21: 3902-3914.

Reviews:

GA Khoury, J Smadbeck, CA Kieslich, CA Floudas (2014) Protein folding and de novo protein design for biotechnological applications. Trends in Biotech., 32 (2), 99-109.

CA Kieslich*, P Tamamis*, RD Gorham Jr, A López de Victoria, NU Sausman, G Archontis, D Morikis (2012) Exploring protein-protein and protein-ligand interactions in the immune system using molecular dynamics and continuum electrostatics. Curr. Phys. Chem., 20(2):324-343.

RD Gorham Jr, CA Kieslich, D Morikis (2011) Electrostatic clustering and free energy calculations provide a foundation for protein design and optimization. Ann. Biomed Eng, 39(4): 1252-63.