@article{SechiFackeldeyChewleetal.2022, author = {Sechi, Renata and Fackeldey, Konstantin and Chewle, Surahit and Weber, Marcus}, title = {SepFree NMF: A Toolbox for Analyzing the Kinetics of Sequential Spectroscopic Data}, series = {Algorithms}, volume = {15}, journal = {Algorithms}, number = {9}, doi = {10.3390/a15090297}, pages = {297}, year = {2022}, abstract = {This work addresses the problem of determining the number of components from sequential spectroscopic data analyzed by non-negative matrix factorization without separability assumption (SepFree NMF). These data are stored in a matrix M of dimension "measured times" versus "measured wavenumbers" and can be decomposed to obtain the spectral fingerprints of the states and their evolution over time. SepFree NMF assumes a memoryless (Markovian) process to underline the dynamics and decomposes M so that M=WH, with W representing the components' fingerprints and H their kinetics. However, the rank of this decomposition (i.e., the number of physical states in the process) has to be guessed from pre-existing knowledge on the observed process. We propose a measure for determining the number of components with the computation of the minimal memory effect resulting from the decomposition; by quantifying how much the obtained factorization is deviating from the Markovian property, we are able to score factorizations of a different number of components. In this way, we estimate the number of different entities which contribute to the observed system, and we can extract kinetic information without knowing the characteristic spectra of the single components. This manuscript provides the mathematical background as well as an analysis of computer generated and experimental sequentially measured Raman spectra.}, language = {en} } @misc{WitzigBeckenbachEifleretal., author = {Witzig, Jakob and Beckenbach, Isabel and Eifler, Leon and Fackeldey, Konstantin and Gleixner, Ambros and Grever, Andreas and Weber, Marcus}, title = {Mixed-Integer Programming for Cycle Detection in Non-reversible Markov Processes}, issn = {1438-0064}, doi = {10.1137/16M1091162}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-60353}, abstract = {In this paper, we present a new, optimization-based method to exhibit cyclic behavior in non-reversible stochastic processes. While our method is general, it is strongly motivated by discrete simulations of ordinary differential equations representing non-reversible biological processes, in particular molecular simulations. Here, the discrete time steps of the simulation are often very small compared to the time scale of interest, i.e., of the whole process. In this setting, the detection of a global cyclic behavior of the process becomes difficult because transitions between individual states may appear almost reversible on the small time scale of the simulation. We address this difficulty using a mixed-integer programming model that allows us to compute a cycle of clusters with maximum net flow, i.e., large forward and small backward probability. For a synthetic genetic regulatory network consisting of a ring-oscillator with three genes, we show that this approach can detect the most productive overall cycle, outperforming classical spectral analysis methods. Our method applies to general non-equilibrium steady state systems such as catalytic reactions, for which the objective value computes the effectiveness of the catalyst.}, language = {en} } @article{WitzigBeckenbachEifleretal., author = {Witzig, Jakob and Beckenbach, Isabel and Eifler, Leon and Fackeldey, Konstantin and Gleixner, Ambros and Grever, Andreas and Weber, Marcus}, title = {Mixed-Integer Programming for Cycle Detection in Non-reversible Markov Processes}, series = {Multiscale Modeling and Simulation}, volume = {16}, journal = {Multiscale Modeling and Simulation}, number = {1}, issn = {1438-0064}, doi = {10.1137/16M1091162}, pages = {248 -- 265}, abstract = {In this paper, we present a new, optimization-based method to exhibit cyclic behavior in non-reversible stochastic processes. While our method is general, it is strongly motivated by discrete simulations of ordinary differential equations representing non-reversible biological processes, in particular molecular simulations. Here, the discrete time steps of the simulation are often very small compared to the time scale of interest, i.e., of the whole process. In this setting, the detection of a global cyclic behavior of the process becomes difficult because transitions between individual states may appear almost reversible on the small time scale of the simulation. We address this difficulty using a mixed-integer programming model that allows us to compute a cycle of clusters with maximum net flow, i.e., large forward and small backward probability. For a synthetic genetic regulatory network consisting of a ring-oscillator with three genes, we show that this approach can detect the most productive overall cycle, outperforming classical spectral analysis methods. Our method applies to general non-equilibrium steady state systems such as catalytic reactions, for which the objective value computes the effectiveness of the catalyst.}, language = {en} } @inproceedings{FackeldeyBujotzek, author = {Fackeldey, Konstantin and Bujotzek, Alexander}, title = {Local Quantum-Like Updates in Classical Molecular Simulation Realized Within an Uncoupling-Coupling Approach}, series = {Progress in Industrial Mathematics at ECMI 2012}, volume = {19}, booktitle = {Progress in Industrial Mathematics at ECMI 2012}, doi = {10.1007/978-3-319-05365-3_42}, pages = {309 -- 313}, abstract = {In this article a method to improve the precision of the classical molecular dynamics force field by solving an approximation problem with scattered quantum mechanical data is presented. This novel technique is based on two steps. In the first step a partition of unity scheme is used for partitioning the state space by meshfree basis functions. As a consequence the potential can be localized for each basis function. In a second step, for one state in each meshfree basis function, the precise QM-based charges are computed. These local QM-based charges are then used, to optimize the local potential function. The performance of this method is shown for the alanine tripeptide.}, language = {en} } @article{RoehlWeberFackeldey, author = {R{\"o}hl, Susanne and Weber, Marcus and Fackeldey, Konstantin}, title = {Computing the minimal rebinding effect for non-reversible processes}, series = {Multiscale Modeling and Simulation}, volume = {19}, journal = {Multiscale Modeling and Simulation}, number = {1}, doi = {https://doi.org/10.1137/20M1334966}, pages = {460 -- 477}, abstract = {The aim of this paper is to investigate the rebinding effect, a phenomenon describing a "short-time memory" which can occur when projecting a Markov process onto a smaller state space. For guaranteeing a correct mapping by the Markov State Model, we assume a fuzzy clustering in terms of membership functions, assigning degrees of membership to each state. The macro states are represented by the membership functions and may be overlapping. The magnitude of this overlap is a measure for the strength of the rebinding effect, caused by the projection and stabilizing the system. A minimal bound for the rebinding effect included in a given system is computed as the solution of an optimization problem. Based on membership functions chosen as a linear combination of Schur vectors, this generalized approach includes reversible as well as non-reversible processes.}, language = {en} } @misc{WeberFackeldey, author = {Weber, Marcus and Fackeldey, Konstantin}, title = {G-PCCA: Spectral Clustering for Non-reversible Markov Chains}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-55505}, abstract = {Spectral clustering methods are based on solving eigenvalue problems for the identification of clusters, e.g., the identification of metastable subsets of a Markov chain. Usually, real-valued eigenvectors are mandatory for this type of algorithms. The Perron Cluster Analysis (PCCA+) is a well-known spectral clustering method of Markov chains. It is applicable for reversible Markov chains, because reversibility implies a real-valued spectrum. We extend this spectral clustering method also to non-reversible Markov chains and give some illustrative examples. The main idea is to replace the eigenvalue problem by a real-valued Schur decomposition. By this extension, non-reversible Markov chains can be analyzed. Furthermore, the chains need not have a positive stationary distribution. And additionally to metastabilities, dominant cycles and sinks can be identified, too.}, language = {en} } @article{ReuterFackeldeyWeber, author = {Reuter, Bernhard and Fackeldey, Konstantin and Weber, Marcus}, title = {Generalized Markov modeling of nonreversible molecular kinetics}, series = {The Journal of Chemical Physics}, volume = {17}, journal = {The Journal of Chemical Physics}, number = {150}, doi = {10.1063/1.5064530}, pages = {174103}, abstract = {Markov state models are to date the gold standard for modeling molecular kinetics since they enable the identification and analysis of metastable states and related kinetics in a very instructive manner. The state-of-the-art Markov state modeling methods and tools are very well developed for the modeling of reversible processes in closed equilibrium systems. On the contrary, they are largely not well suited to deal with nonreversible or even nonautonomous processes of nonequilibrium systems. Thus, we generalized the common Robust Perron Cluster Cluster Analysis (PCCA+) method to enable straightforward modeling of nonequilibrium systems as well. The resulting Generalized PCCA (G-PCCA) method readily handles equilibrium as well as nonequilibrium data by utilizing real Schur vectors instead of eigenvectors. This is implemented in the G-PCCA algorithm that enables the semiautomatic coarse graining of molecular kinetics. G-PCCA is not limited to the detection of metastable states but also enables the identification and modeling of cyclic processes. This is demonstrated by three typical examples of nonreversible systems.}, language = {en} } @inproceedings{FackeldeyBujotzekWeber, author = {Fackeldey, Konstantin and Bujotzek, Alexander and Weber, Marcus}, title = {A meshless discretization method for Markov state models applied to explicit water peptide folding simulations}, series = {Meshfree Methods for Partial Differential Equations VI}, volume = {89}, booktitle = {Meshfree Methods for Partial Differential Equations VI}, publisher = {Springer}, pages = {141 -- 154}, language = {en} } @misc{BujotzekSchuettNielsenetal., author = {Bujotzek, Alexander and Sch{\"u}tt, Ole and Nielsen, Adam and Fackeldey, Konstantin and Weber, Marcus}, title = {Efficient Conformational Analysis by Partition-of-Unity Coupling}, series = {Math Chem}, journal = {Math Chem}, issn = {1438-0064}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-42570}, abstract = {Obtaining a sufficient sampling of conformational space is a common problem in molecular simulation. We present the implementation of an umbrella-like adaptive sampling approach based on function-based meshless discretization of conformational space that is compatible with state of the art molecular dynamics code and that integrates an eigenvector-based clustering approach for conformational analysis and the computation of inter-conformational transition rates. The approach is applied to three example systems, namely n-pentane, alanine dipeptide, and a small synthetic host-guest system, the latter two including explicitly modeled solvent.}, language = {en} } @misc{FackeldeySikorskiWeber, author = {Fackeldey, Konstantin and Sikorski, Alexander and Weber, Marcus}, title = {Spectral Clustering for Non-reversible Markov Chains}, issn = {1438-0064}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-70218}, abstract = {Spectral clustering methods are based on solving eigenvalue problems for the identification of clusters, e.g. the identification of metastable subsets of a Markov chain. Usually, real-valued eigenvectors are mandatory for this type of algorithms. The Perron Cluster Analysis (PCCA+) is a well-known spectral clustering method of Markov chains. It is applicable for reversible Markov chains, because reversibility implies a real-valued spectrum. We also extend this spectral clustering method to non-reversible Markov chains and give some illustrative examples. The main idea is to replace the eigenvalue problem by a real-valued Schur decomposition. By this extension non-reversible Markov chains can be analyzed. Furthermore, the chains do not need to have a positive stationary distribution. In addition to metastabilities, dominant cycles and sinks can also be identified. This novel method is called GenPCCA (i.e. Generalized PCCA), since it includes the case of non reversible processes. We also apply the method to real world eye tracking data.}, language = {en} } @article{WeberFackeldeySchuette, author = {Weber, Marcus and Fackeldey, Konstantin and Sch{\"u}tte, Christof}, title = {Set-Free Markov State Model Building}, series = {Journal of Chemical Physics}, volume = {146}, journal = {Journal of Chemical Physics}, number = {12}, doi = {10.1063/1.4978501}, language = {en} } @phdthesis{Fackeldey, author = {Fackeldey, Konstantin}, title = {Crossing the Scales in Structural Mechanics and Molecular Research}, language = {en} } @misc{Fackeldey, author = {Fackeldey, Konstantin}, title = {Coupling Meshbased and Meshfree Methods by a Transfer Operator Approach}, issn = {1438-0064}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-11755}, number = {10-12}, abstract = {In contrast to the well known meshbased methods like the finite element method, meshfree methods do not rely on a mesh. However besides their great applicability, meshfree methods are rather time consuming. Thus, it seems favorable to combine both methods, by using meshfree methods only in a small part of the domain, where a mesh is disadvantageous, and a meshbased method for the rest of the domain. We motivate, that this coupling between the two simulation techniques can be considered as saddle point problem and show the stability of this coupling. Thereby a novel transfer operator is introduced, which interacts in the transition zone, where both methods coexist.}, language = {en} } @misc{FackeldeyRoeblitzScharkoietal., author = {Fackeldey, Konstantin and R{\"o}blitz, Susanna and Scharkoi, Olga and Weber, Marcus}, title = {Soft Versus Hard Metastable Conformations in Molecular Simulations}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-13189}, number = {11-27}, abstract = {Particle methods have become indispensible in conformation dynamics to compute transition rates in protein folding, binding processes and molecular design, to mention a few. Conformation dynamics requires at a decomposition of a molecule's position space into metastable conformations. In this paper, we show how this decomposition can be obtained via the design of either ``soft'' or ``hard'' molecular conformations. We show, that the soft approach results in a larger metastabilitiy of the decomposition and is thus more advantegous. This is illustrated by a simulation of Alanine Dipeptide.}, language = {en} } @article{WeberFackeldey, author = {Weber, Marcus and Fackeldey, Konstantin}, title = {Local Refinements in Classical Molecular Dynamics Simulations}, series = {J. Phys. Conf. Ser.}, volume = {490}, journal = {J. Phys. Conf. Ser.}, pages = {012016}, language = {en} } @article{GorgullaDasLeighetal., author = {Gorgulla, Christoph and Das, Krishna M. Padmanabha and Leigh, Kendra E and Cespugli, Marco and Fischer, Patrick D. and Wang, Zi-Fu and Tesseyre, Guilhem and Pandita, Shreya and Shnapir, Alex and Calderaio, Anthony and Hutcheson, Colin and Gechev, Minko and Rose, Alexander and Lewis, Noam and Yaffe, Erez and Luxenburg, Roni and Herce, Henry D. and Durmaz, Vedat and Halazonetis, Thanos D. and Fackeldey, Konstantin and Patten, Justin J. and Chuprina, Alexander and Dziuba, Igor and Plekhova, Alla and Moroz, Yurii and Radchenko, Dmytro and Tarkhanova, Olga and Yavnyuk, Irina and Gruber, Christian C. and Yust, Ryan and Payne, Dave and N{\"a}{\"a}r, Anders M. and Namchuk, Mark N. and Davey, Robert A. and Wagner, Gerhard and Kinney, Jamie and Arthanari, Haribabu}, title = {A Multi-Pronged Approach Targeting SARS-CoV-2 Proteins Using Ultra-Large Virtual Screening}, series = {iScience}, volume = {24}, journal = {iScience}, number = {2}, publisher = {CellPress}, doi = {10.26434/chemrxiv.12682316}, pages = {102021}, abstract = {Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), previously known as 2019 novel coronavirus (2019-nCoV), has spread rapidly across the globe, creating an unparalleled global health burden and spurring a deepening economic crisis. As of July 7th, 2020, almost seven months into the outbreak, there are no approved vaccines and few treatments available. Developing drugs that target multiple points in the viral life cycle could serve as a strategy to tackle the current as well as future coronavirus pandemics. Here we leverage the power of our recently developed in silico screening platform, VirtualFlow, to identify inhibitors that target SARS-CoV-2. VirtualFlow is able to efficiently harness the power of computing clusters and cloud-based computing platforms to carry out ultra-large scale virtual screens. In this unprecedented structure-based multi-target virtual screening campaign, we have used VirtualFlow to screen an average of approximately 1 billion molecules against each of 40 different target sites on 17 different potential viral and host targets in the cloud. In addition to targeting the active sites of viral enzymes, we also target critical auxiliary sites such as functionally important protein-protein interaction interfaces. This multi-target approach not only increases the likelihood of finding a potent inhibitor, but could also help identify a collection of anti-coronavirus drugs that would retain efficacy in the face of viral mutation. Drugs belonging to different regimen classes could be combined to develop possible combination therapies, and top hits that bind at highly conserved sites would be potential candidates for further development as coronavirus drugs. Here, we present the top 200 in silico hits for each target site. While in-house experimental validation of some of these compounds is currently underway, we want to make this array of potential inhibitor candidates available to researchers worldwide in consideration of the pressing need for fast-tracked drug development.}, language = {en} } @article{GorgullaBoeszoermnyiWangetal., author = {Gorgulla, Christoph and Boeszoermnyi, Andras and Wang, Zi-Fu and Fischer, Patrick D. and Coote, Paul and Das, Krishna M. Padmanabha and Malets, Yehor S. and Radchenko, Dmytro S. and Moroz, Yurii and Scott, David A. and Fackeldey, Konstantin and Hoffmann, Moritz and Iavniuk, Iryna and Wagner, Gerhard and Arthanari, Haribabu}, title = {An open-source drug discovery platform enables ultra-large virtual screens}, series = {Nature}, volume = {580}, journal = {Nature}, publisher = {Springer Nature}, doi = {https://doi.org/10.1038/s41586-020-2117-z}, pages = {663 -- 668}, abstract = {On average, an approved drug today costs \$2-3 billion and takes over ten years to develop1. In part, this is due to expensive and time-consuming wet-lab experiments, poor initial hit compounds, and the high attrition rates in the (pre-)clinical phases. Structure-based virtual screening (SBVS) has the potential to mitigate these problems. With SBVS, the quality of the hits improves with the number of compounds screened2. However, despite the fact that large compound databases exist, the ability to carry out large-scale SBVSs on computer clusters in an accessible, efficient, and flexible manner has remained elusive. Here we designed VirtualFlow, a highly automated and versatile open-source platform with perfect scaling behaviour that is able to prepare and efficiently screen ultra-large ligand libraries of compounds. VirtualFlow is able to use a variety of the most powerful docking programs. Using VirtualFlow, we have prepared the largest and freely available ready-to-dock ligand library available, with over 1.4 billion commercially available molecules. To demonstrate the power of VirtualFlow, we screened over 1 billion compounds and discovered a small molecule inhibitor (iKeap1) that engages KEAP1 with nanomolar affinity (Kd = 114 nM) and disrupts the interaction between KEAP1 and the transcription factor NRF2. We also identified a set of structurally diverse molecules that bind to KEAP1 with submicromolar affinity. This illustrates the potential of VirtualFlow to access vast regions of the chemical space and identify binders with high affinity for target proteins.}, language = {en} } @article{GorgullaCınaroğluFischeretal., author = {Gorgulla, Christoph and {\c{C}}{\i}naroğlu, S{\"u}leyman and Fischer, Patrick D. and Fackeldey, Konstantin and Wagner, Gerhard and Arthanari, Haribabu}, title = {VirtualFlow Ants—Ultra-Large Virtual Screenings with Artificial Intelligence Driven Docking Algorithm Based on Ant Colony Optimization}, series = {Special Issue Artificial Intelligence \& Deep Learning Approaches for Structural Bioinformatics}, volume = {22}, journal = {Special Issue Artificial Intelligence \& Deep Learning Approaches for Structural Bioinformatics}, number = {11}, doi = {https://doi.org/10.3390/ijms22115807}, pages = {5807}, abstract = {The docking program PLANTS, which is based on ant colony optimization (ACO) algorithm, has many advanced features for molecular docking. Among them are multiple scoring functions, the possibility to model explicit displaceable water molecules, and the inclusion of experimental constraints. Here, we add support of PLANTS to VirtualFlow (VirtualFlow Ants), which adds a valuable method for primary virtual screenings and rescoring procedures. Furthermore, we have added support of ligand libraries in the MOL2 format, as well as on the fly conversion of ligand libraries which are in the PDBQT format to the MOL2 format to endow VirtualFlow Ants with an increased flexibility regarding the ligand libraries. The on the fly conversion is carried out with Open Babel and the program SPORES. We applied VirtualFlow Ants to a test system involving KEAP1 on the Google Cloud up to 128,000 CPUs, and the observed scaling behavior is approximately linear. Furthermore, we have adjusted several central docking parameters of PLANTS (such as the speed parameter or the number of ants) and screened 10 million compounds for each of the 10 resulting docking scenarios. We analyzed their docking scores and average docking times, which are key factors in virtual screenings. The possibility of carrying out ultra-large virtual screening with PLANTS via VirtualFlow Ants opens new avenues in computational drug discovery.}, language = {en} } @article{FackeldeySikorskiWeber, author = {Fackeldey, Konstantin and Sikorski, Alexander and Weber, Marcus}, title = {Spectral Clustering for Non-Reversible Markov Chains}, series = {Computational and Applied Mathematics}, volume = {37}, journal = {Computational and Applied Mathematics}, number = {5}, doi = {https://doi.org/10.1007/s40314-018-0697-0}, pages = {6376 -- 6391}, abstract = {Spectral clustering methods are based on solving eigenvalue problems for the identification of clusters, e.g., the identification of metastable subsets of a Markov chain. Usually, real-valued eigenvectors are mandatory for this type of algorithms. The Perron Cluster Analysis (PCCA+) is a well-known spectral clustering method of Markov chains. It is applicable for reversible Markov chains, because reversibility implies a real-valued spectrum. We also extend this spectral clustering method to non-reversible Markov chains and give some illustrative examples. The main idea is to replace the eigenvalue problem by a real-valued Schur decomposition. By this extension non-reversible Markov chains can be analyzed. Furthermore, the chains do not need to have a positive stationary distribution. In addition to metastabilities, dominant cycles and sinks can also be identified. This novel method is called GenPCCA (i.e., generalized PCCA), since it includes the case of non-reversible processes. We also apply the method to real-world eye-tracking data.}, language = {en} } @article{ErnstFackeldeyVolkameretal., author = {Ernst, Natalia and Fackeldey, Konstantin and Volkamer, Andrea and Opatz, Oliver and Weber, Marcus}, title = {Computation of temperature-dependent dissociation rates of metastable protein-ligand complexes}, series = {Molecular Simulation}, volume = {45}, journal = {Molecular Simulation}, number = {11}, doi = {10.1080/08927022.2019.1610949}, pages = {904 -- 911}, abstract = {Molecular simulations are often used to analyse the stability of protein-ligand complexes. The stability can be characterised by exit rates or using the exit time approach, i.e. by computing the expected holding time of the complex before its dissociation. However determining exit rates by straightforward molecular dynamics methods can be challenging for stochastic processes in which the exit event occurs very rarely. Finding a low variance procedure for collecting rare event statistics is still an open problem. In this work we discuss a novel method for computing exit rates which uses results of Robust Perron Cluster Analysis (PCCA+). This clustering method gives the possibility to define a fuzzy set by a membership function, which provides additional information of the kind 'the process is being about to leave the set'. Thus, the derived approach is not based on the exit event occurrence and, therefore, is also applicable in case of rare events. The novel method can be used to analyse the temperature effect of protein-ligand systems through the differences in exit rates, and, thus, open up new drug design strategies and therapeutic applications.}, language = {en} }