Automated Deep Abstractions for Stochastic Chemical Reaction Networks

Predicting stochastic cellular dynamics as emerging from the mechanistic models of molecular interactions is a long-standing challenge in systems biology: low-level chemical reaction network (CRN) models give raise to a highly-dimensional continuous-time Markov chain (CTMC) which is computationally demanding and often prohibitive to analyse in practice. A recently proposed abstraction method uses deep learning to replace this CTMC with a discrete-time continuous-space process, by training a mixture density deep neural network with traces sampled at regular time intervals (which can obtained either by simulating a given CRN or as time-series data from experiment). The major advantage of such abstraction is that it produces a computational model that is dramatically cheaper to execute, while preserving the statistical features of the training data. In general, the abstraction accuracy improves with the amount of training data. However, depending on a CRN, the overall quality of the method -- the efficiency gain and abstraction accuracy -- will also depend on the choice of neural network architecture given by hyper-parameters such as the layer types and connections between them. As a consequence, in practice, the modeller would have to take care of finding the suitable architecture manually, for each given CRN, through a tedious and time-consuming trial-and-error cycle. In this paper, we propose to further automatise deep abstractions for stochastic CRNs, through learning the optimal neural network architecture along with learning the transition kernel of the abstract process. Automated search of the architecture makes the method applicable directly to any given CRN, which is time-saving for deep learning experts and crucial for non-specialists. We implement the method and demonstrate its performance on a number of representative CRNs with multi-modal emergent phenotypes.

[1]  D. Gillespie Exact Stochastic Simulation of Coupled Chemical Reactions , 1977 .

[2]  Quoc V. Le,et al.  Neural Architecture Search with Reinforcement Learning , 2016, ICLR.

[3]  Christopher N Davis,et al.  The use of mixture density networks in the emulation of complex epidemiological individual-based models , 2019, bioRxiv.

[4]  Heinz Koeppl,et al.  Markov chain aggregation and its applications to combinatorial reaction networks , 2014, Journal of mathematical biology.

[5]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[6]  Andreea Beica,et al.  Efficient Reduction of Kappa Models by Static Inspection of the Rule-Set , 2015, HSB.

[7]  Heinz Koeppl,et al.  Stochastic Fragments: A Framework for the Exact Reduction of the Stochastic Semantics of Rule-Based Models , 2013, Int. J. Softw. Informatics.

[8]  Heinz Koeppl,et al.  Lumpability abstractions of rule-based systems , 2010, Theor. Comput. Sci..

[9]  C. Bishop Mixture density networks , 1994 .

[10]  Yoshua Bengio,et al.  BinaryConnect: Training Deep Neural Networks with binary weights during propagations , 2015, NIPS.

[11]  Linda R. Petzold,et al.  Accuracy limitations and the measurement of errors in the stochastic simulation of chemically reacting systems , 2006, J. Comput. Phys..

[12]  David F. Anderson,et al.  Continuous Time Markov Chain Models for Chemical Reaction Networks , 2011 .

[13]  Tatjana Petrov,et al.  StochNetV2: A Tool for Automated Deep Abstractions for Stochastic Reaction Networks , 2020, QEST.

[14]  David Safránek,et al.  Data-Informed Parameter Synthesis for Population Markov Chains , 2019, HSB.

[15]  Guido Sanguinetti,et al.  Statistical Abstraction for Multi-scale Spatio-temporal Systems , 2019, ACM Trans. Model. Comput. Simul..

[16]  RADEK ERBAN,et al.  Noise-induced mixing and multimodality in reaction networks , 2018, European Journal of Applied Mathematics.

[17]  Andrea Vandin,et al.  SPEEDING UP STOCHASTIC AND DETERMINISTIC SIMULATION BY AGGREGATION: AN ADVANCED TUTORIAL , 2018, 2018 Winter Simulation Conference (WSC).

[18]  Alberto Policriti,et al.  On the impact of discreteness and abstractions on modelling noise in gene regulatory networks , 2015, Comput. Biol. Chem..

[19]  Julio Saez-Rodriguez,et al.  A domain-oriented approach to the reduction of combinatorial complexity in signal transduction networks , 2006, BMC Bioinformatics.

[20]  Guido Sanguinetti,et al.  Statistical Abstraction for Multi-scale Spatio-temporal Systems , 2017, QEST.

[21]  Luca Cardelli,et al.  Syntactic Markovian Bisimulation for Chemical Reaction Networks , 2017, Models, Algorithms, Logics and Tools.

[22]  Chetan D. Pahlajani,et al.  Stochastic reduction method for biological chemical kinetics using time-scale separation. , 2011, Journal of theoretical biology.

[23]  Yiming Yang,et al.  DARTS: Differentiable Architecture Search , 2018, ICLR.

[24]  Thomas A. Henzinger,et al.  Hybrid numerical solution of the chemical master equation , 2010, CMSB '10.

[25]  Jeremy Gunawardena,et al.  A Linear Framework for Time-Scale Separation in Nonlinear Biochemical Systems , 2012, PloS one.

[26]  Bastien Chopard,et al.  Multiscale modelling and simulation: a position paper , 2014, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences.

[27]  Song Han,et al.  ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware , 2018, ICLR.

[28]  Matthieu Pichené,et al.  Abstracting the dynamics of biological pathways using information theory: a case study of apoptosis pathway , 2017, Bioinform..

[29]  Quoc V. Le,et al.  Searching for Activation Functions , 2018, arXiv.

[30]  Luca Bortolussi,et al.  Deep Abstractions of Chemical Reaction Networks , 2018, CMSB.

[31]  T. Kurtz Limit theorems for sequences of jump Markov processes approximating ordinary differential processes , 1971, Journal of Applied Probability.

[32]  Elijah Roberts,et al.  Approximation and inference methods for stochastic biochemical kinetics—a tutorial review , 2017 .

[33]  Luca Cardelli,et al.  Stochastic analysis of Chemical Reaction Networks using Linear Noise Approximation. , 2016, Bio Systems.