Analyzing protein dynamics from fluorescence intensity traces using unsupervised deep learning network

We propose an unsupervised deep learning network to analyze the dynamics of membrane proteins from the fluorescence intensity traces. This system was trained in an unsupervised manner with the raw experimental time traces and synthesized ones, so neither predefined state number nor pre-labelling were required. With the bidirectional Long Short-Term Memory (biLSTM) networks as the hidden layers, both the past and future context can be used fully to improve the prediction results and can even extract information from the noise distribution. The method was validated with the synthetic dataset and the experimental dataset of monomeric fluorophore Cy5, and then applied to extract the membrane protein interaction dynamics from experimental data successfully. Yuan et al. propose an unsupervised deep learning network approach to analyze the dynamics of membrane proteins from the fluorescence intensity traces. The unsupervised nature facilitates training of the system without predefined state number or pre-labelling and can even extract information from noise distribution.

[1]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[2]  Taekjip Ha,et al.  Probing Cellular Protein Complexes via Single Molecule Pull-down , 2011, Nature.

[3]  Jürgen Schmidhuber,et al.  Framewise phoneme classification with bidirectional LSTM and other neural network architectures , 2005, Neural Networks.

[4]  Adrian Whitty,et al.  Understanding cytokine and growth factor receptor activation mechanisms , 2012, Critical reviews in biochemistry and molecular biology.

[5]  Hossein Jashnsaz,et al.  A novel method to accurately locate and count large numbers of steps by photobleaching , 2016, Molecular biology of the cell.

[6]  R. A. Kennedy,et al.  Forward-backward non-linear filtering technique for extracting small biological signals from noise , 1991, Journal of Neuroscience Methods.

[7]  X. Fang,et al.  Analysis of the steps in single-molecule photobleaching traces by using the hidden markov model and maximum-likelihood clustering. , 2014, Chemistry, an Asian journal.

[8]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[9]  Derek Bowie,et al.  Automating Single Subunit Counting of Membrane Proteins in Mammalian Cells* , 2012, The Journal of Biological Chemistry.

[10]  E. Isacoff,et al.  Stoichiometry of the KCNQ1 - KCNE1 ion channel complex , 2010, Proceedings of the National Academy of Sciences.

[11]  Kaizhu Huang,et al.  Generative adversarial networks with decoder-encoder output noises , 2020, Neural Networks.

[12]  A. Ng.  Feature selection, L1 vs. L2 regularization, and rotational invariance , 2004, Twenty-first international conference on Machine learning - ICML '04.

[13]  Nasser M. Nasrabadi,et al.  Pattern Recognition and Machine Learning , 2006, Technometrics.

[14]  Taekjip Ha,et al.  Single-molecule methods leap ahead , 2014, Nature Methods.

[15]  Liedewij Laan,et al.  Assembly dynamics of microtubules at molecular resolution , 2006, Nature.

[16]  A. Sergé,et al.  Dynamic multiple-target tracing to probe spatiotemporal cartography of cell membranes , 2008, Nature Methods.

[17]  E. Isacoff,et al.  Subunit counting in membrane-bound proteins , 2007, Nature Methods.

[18]  X. Fang,et al.  Monomeric type I and type III transforming growth factor-β receptors and their dimerization revealed by single-molecule imaging , 2010, Cell Research.

[19]  T. Thireou,et al.  Bidirectional Long Short-Term Memory Networks for Predicting the Subcellular Localization of Eukaryotic Proteins , 2007, IEEE/ACM Transactions on Computational Biology and Bioinformatics.

[20]  Gaudenz Danuser,et al.  Cytoskeletal Control of CD36 Diffusion Promotes Its Receptor and Signaling Function , 2011, Cell.

[21]  H. Bayley,et al.  Membrane Protein Stoichiometry Determined from the Step‐Wise Photobleaching of Dye‐Labelled Subunits , 2007, Chembiochem : a European journal of chemical biology.

[22]  Brian M. Sadler,et al.  Analysis of Multiscale Products for Step Detection and Estimation , 1999, IEEE Trans. Inf. Theory.

[23]  Ye Guang Chen,et al.  Lateral diffusion of TGF-β type I receptor studied by single-molecule imaging , 2007 .

[24]  Santo Fortunato,et al.  Weight Thresholding on Complex Networks , 2018, Physical Review E.

[25]  Léon Bottou,et al.  Wasserstein Generative Adversarial Networks , 2017, ICML.

[26]  R. Levy,et al.  Direct Determination of Kinetic Rates from Single-Molecule Photon Arrival Trajectories Using Hidden Markov Models. , 2003, The journal of physical chemistry. A.

[27]  Akihiro Kusumi,et al.  Ultrafast Diffusion of a Fluorescent Cholesterol Analog in Compartmentalized Plasma Membranes , 2014, Traffic.

[28]  Yazhen Wang.  Jump and sharp cusp detection by wavelets , 1995 .

[29]  SchmidhuberJürgen,et al.  2005 Special Issue , 2005 .

[30]  R. Cross,et al.  Mechanics of the kinesin step , 2005, Nature.

[31]  John R. James,et al.  BRET analysis of GPCR oligomerization: newer does not mean better , 2007, Nature Methods.

[32]  Xuedong Liu,et al.  Decoding the quantitative nature of TGF-beta/Smad signaling. , 2008, Trends in cell biology.

[33]  Sean R Eddy,et al.  What is a hidden Markov model? , 2004, Nature Biotechnology.

[34]  W. Zhou,et al.  Single-Molecule Imaging Reveals the Activation Dynamics of Intracellular Protein Smad3 on Cell Membrane , 2016, Scientific Reports.

[35]  Murti Salapaka,et al.  Detection of Steps in Single Molecule Data , 2011, Cellular and Molecular Bioengineering.

[36]  K. Jaqaman,et al.  Robust single particle tracking in live cell time-lapse sequences , 2008, Nature Methods.

[37]  Robert L. Mercer,et al.  The Mathematics of Statistical Machine Translation: Parameter Estimation , 1993, CL.

[38]  Qianqian Wang,et al.  Single-molecule imaging reveals transforming growth factor-β-induced type II receptor dimerization , 2009, Proceedings of the National Academy of Sciences.

[39]  Long Zhang,et al.  A special issue on TGF-β signaling: regulation, crosstalk, and biology. , 2018, Acta biochimica et biophysica Sinica.

[40]  X. Fang,et al.  Single-molecule imaging reveals the stoichiometry change of epidermal growth factor receptor during transactivation by β2-adrenergic receptor , 2017, Science China Chemistry.

[41]  A. Knight,et al.  Coupling ATP hydrolysis to mechanical work , 1999, Nature Cell Biology.

[42]  Xiaohong Fang,et al.  Automated Stoichiometry Analysis of Single-Molecule Fluorescence Imaging Traces via Deep Learning. , 2019, Journal of the American Chemical Society.

[43]  Akihiro Kusumi,et al.  Tracking single molecules at work in living cells. , 2014, Nature chemical biology.

[44]  Phil Blunsom,et al.  Recurrent Continuous Translation Models , 2013, EMNLP.

[45]  X. Fang,et al.  Quantitative single-molecule study of TGF-&bgr;/Smad signaling , 2018, Acta biochimica et biophysica Sinica.

[46]  Paul A. Gagniuc,et al.  Markov Chains: From Theory to Implementation and Experimentation , 2017 .

[47]  A. Hinck,et al.  TGF‐β signalling is mediated by two autonomously functioning TβRI:TβRII pairs , 2011, The EMBO journal.

[48]  J. Elf,et al.  Extracting intracellular diffusive states and transition rates from single-molecule tracking data , 2013, Nature Methods.

[49]  Troy C Messina,et al.  Hidden Markov model analysis of multichromophore photobleaching. , 2006, The journal of physical chemistry. B.

[50]  Jin-Ho Ahn,et al.  NoRSE: noise reduction and state evaluator for high-frequency single event traces , 2012, Bioinform..