1
|
Abstract
Gaussian process dynamical systems (GPDS) represent Bayesian nonparametric approaches to inference of nonlinear dynamical systems, and provide a principled framework for the learning of biological networks from multiple perturbed time series measurements of gene or protein expression. Such approaches are able to capture the full richness of complex ODE models, and can be scaled for inference in moderately large systems containing hundreds of genes. Related hierarchical approaches allow for inference from multiple datasets in which the underlying generative networks are assumed to have been rewired, either by context-dependent changes in network structure, evolutionary processes, or synthetic manipulation. These approaches can also be used to leverage experimentally determined network structures from one species into another where the network structure is unknown. Collectively, these methods provide a comprehensive and flexible platform for inference from a diverse range of data, with applications in systems and synthetic biology, as well as spatiotemporal modelling of embryo development. In this chapter we provide an overview of GPDS approaches and highlight their applications in the biological sciences, with accompanying tutorials available as a Jupyter notebook from https://github.com/cap76/GPDS .
Collapse
Affiliation(s)
| | - Iulia Gherman
- Warwick Integrative Synthetic Biology Centre, School of Engineering, University of Warwick, Coventry, UK
| | - Anastasiya Sybirna
- Wellcome/CRUK Gurdon Institute, University of Cambridge, Cambridge, UK
- Wellcome/MRC Cambridge Stem Cell Institute, University of Cambridge, Cambridge, UK
- Physiology, Development and Neuroscience Department, University of Cambridge, Cambridge, UK
| | - David L Wild
- Department of Statistics and Systems Biology Centre, University of Warwick, Coventry, UK
| |
Collapse
|
2
|
Shahdoust M, Pezeshk H, Mahjub H, Sadeghi M. F-MAP: A Bayesian approach to infer the gene regulatory network using external hints. PLoS One 2017; 12:e0184795. [PMID: 28938012 PMCID: PMC5609748 DOI: 10.1371/journal.pone.0184795] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.4] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/22/2017] [Accepted: 08/31/2017] [Indexed: 01/07/2023] Open
Abstract
The Common topological features of related species gene regulatory networks suggest reconstruction of the network of one species by using the further information from gene expressions profile of related species. We present an algorithm to reconstruct the gene regulatory network named; F-MAP, which applies the knowledge about gene interactions from related species. Our algorithm sets a Bayesian framework to estimate the precision matrix of one species microarray gene expressions dataset to infer the Gaussian Graphical model of the network. The conjugate Wishart prior is used and the information from related species is applied to estimate the hyperparameters of the prior distribution by using the factor analysis. Applying the proposed algorithm on six related species of drosophila shows that the precision of reconstructed networks is improved considerably compared to the precision of networks constructed by other Bayesian approaches.
Collapse
Affiliation(s)
- Maryam Shahdoust
- Department of Biostatistics, School of Public Health, Hamadan University of Medical Sciences, Hamadan, Iran
| | - Hamid Pezeshk
- School of Mathematics, Statistics and Computer Science, College of Science, University of Tehran, Tehran, Iran
| | - Hossein Mahjub
- Department of Biostatistics, School of Public Health, Hamadan University of Medical Sciences, Hamadan, Iran
| | - Mehdi Sadeghi
- National Institute of Genetic Engineering and Biotechnology, Tehran, Iran
| |
Collapse
|
3
|
Fused Regression for Multi-source Gene Regulatory Network Inference. PLoS Comput Biol 2016; 12:e1005157. [PMID: 27923054 PMCID: PMC5140053 DOI: 10.1371/journal.pcbi.1005157] [Citation(s) in RCA: 28] [Impact Index Per Article: 3.5] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/22/2016] [Accepted: 09/20/2016] [Indexed: 12/03/2022] Open
Abstract
Understanding gene regulatory networks is critical to understanding cellular differentiation and response to external stimuli. Methods for global network inference have been developed and applied to a variety of species. Most approaches consider the problem of network inference independently in each species, despite evidence that gene regulation can be conserved even in distantly related species. Further, network inference is often confined to single data-types (single platforms) and single cell types. We introduce a method for multi-source network inference that allows simultaneous estimation of gene regulatory networks in multiple species or biological processes through the introduction of priors based on known gene relationships such as orthology incorporated using fused regression. This approach improves network inference performance even when orthology mapping and conservation are incomplete. We refine this method by presenting an algorithm that extracts the true conserved subnetwork from a larger set of potentially conserved interactions and demonstrate the utility of our method in cross species network inference. Last, we demonstrate our method’s utility in learning from data collected on different experimental platforms. Gene regulatory networks describing related biological processes are thought to share conserved interaction structure. This assumption motivates a great deal of work in model systems–where discovery of gene regulation may be more experimentally tractable–but is difficult to directly evaluate using existing methods. The presence of shared structure in a well studied model system or process should make the problem of network inference in a related process easier, but this information is not often applied to the discovery of global gene regulatory networks. Further, to be able to successfully translate findings between different organisms, it is important to be able to identify where regulatory structure is different. We provide a method based on penalized fused regression for inferring gene regulatory networks given prior knowledge about the similarity of interactions in each network. This method is demonstrated on synthetic data, and applied to the problem of inferring networks in distantly related bacterial organisms. We then introduce an extension of the method to deal with the condition of uncertainty over the degree of regulatory conservation by simultaneously inferring gene conservation and interaction weights.
Collapse
|
4
|
Kludas J, Arvas M, Castillo S, Pakula T, Oja M, Brouard C, Jäntti J, Penttilä M, Rousu J. Machine Learning of Protein Interactions in Fungal Secretory Pathways. PLoS One 2016; 11:e0159302. [PMID: 27441920 PMCID: PMC4956264 DOI: 10.1371/journal.pone.0159302] [Citation(s) in RCA: 6] [Impact Index Per Article: 0.8] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/15/2016] [Accepted: 06/30/2016] [Indexed: 12/18/2022] Open
Abstract
In this paper we apply machine learning methods for predicting protein interactions in fungal secretion pathways. We assume an inter-species transfer setting, where training data is obtained from a single species and the objective is to predict protein interactions in other, related species. In our methodology, we combine several state of the art machine learning approaches, namely, multiple kernel learning (MKL), pairwise kernels and kernelized structured output prediction in the supervised graph inference framework. For MKL, we apply recently proposed centered kernel alignment and p-norm path following approaches to integrate several feature sets describing the proteins, demonstrating improved performance. For graph inference, we apply input-output kernel regression (IOKR) in supervised and semi-supervised modes as well as output kernel trees (OK3). In our experiments simulating increasing genetic distance, Input-Output Kernel Regression proved to be the most robust prediction approach. We also show that the MKL approaches improve the predictions compared to uniform combination of the kernels. We evaluate the methods on the task of predicting protein-protein-interactions in the secretion pathways in fungi, S.cerevisiae, baker's yeast, being the source, T. reesei being the target of the inter-species transfer learning. We identify completely novel candidate secretion proteins conserved in filamentous fungi. These proteins could contribute to their unique secretion capabilities.
Collapse
Affiliation(s)
- Jana Kludas
- Helsinki Institute for Information Technology HIIT, Department of Computer Science, Aalto University, Espoo, Finland
| | - Mikko Arvas
- VTT Technical Research Centre of Finland, Espoo, Finland
| | | | - Tiina Pakula
- VTT Technical Research Centre of Finland, Espoo, Finland
| | - Merja Oja
- VTT Technical Research Centre of Finland, Espoo, Finland
| | - Céline Brouard
- Helsinki Institute for Information Technology HIIT, Department of Computer Science, Aalto University, Espoo, Finland
| | - Jussi Jäntti
- VTT Technical Research Centre of Finland, Espoo, Finland
| | - Merja Penttilä
- VTT Technical Research Centre of Finland, Espoo, Finland
| | - Juho Rousu
- Helsinki Institute for Information Technology HIIT, Department of Computer Science, Aalto University, Espoo, Finland
| |
Collapse
|
5
|
Penfold CA, Millar JBA, Wild DL. Inferring orthologous gene regulatory networks using interspecies data fusion. Bioinformatics 2015; 31:i97-105. [PMID: 26072515 PMCID: PMC4765882 DOI: 10.1093/bioinformatics/btv267] [Citation(s) in RCA: 14] [Impact Index Per Article: 1.6] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/31/2022] Open
Abstract
Motivation: The ability to jointly learn gene regulatory networks (GRNs) in, or leverage GRNs between related species would allow the vast amount of legacy data obtained in model organisms to inform the GRNs of more complex, or economically or medically relevant counterparts. Examples include transferring information from Arabidopsis thaliana into related crop species for food security purposes, or from mice into humans for medical applications. Here we develop two related Bayesian approaches to network inference that allow GRNs to be jointly inferred in, or leveraged between, several related species: in one framework, network information is directly propagated between species; in the second hierarchical approach, network information is propagated via an unobserved ‘hypernetwork’. In both frameworks, information about network similarity is captured via graph kernels, with the networks additionally informed by species-specific time series gene expression data, when available, using Gaussian processes to model the dynamics of gene expression. Results: Results on in silico benchmarks demonstrate that joint inference, and leveraging of known networks between species, offers better accuracy than standalone inference. The direct propagation of network information via the non-hierarchical framework is more appropriate when there are relatively few species, while the hierarchical approach is better suited when there are many species. Both methods are robust to small amounts of mislabelling of orthologues. Finally, the use of Saccharomyces cerevisiae data and networks to inform inference of networks in the budding yeast Schizosaccharomyces pombe predicts a novel role in cell cycle regulation for Gas1 (SPAC19B12.02c), a 1,3-beta-glucanosyltransferase. Availability and implementation: MATLAB code is available from http://go.warwick.ac.uk/systemsbiology/software/. Contact:d.l.wild@warwick.ac.uk Supplementary information:Supplementary data are available at Bioinformatics online.
Collapse
Affiliation(s)
- Christopher A Penfold
- Warwick Systems Biology Centre and Biomedical Cell Biology, Warwick Medical School, University of Warwick, Coventry CV4 7AL, UK
| | - Jonathan B A Millar
- Warwick Systems Biology Centre and Biomedical Cell Biology, Warwick Medical School, University of Warwick, Coventry CV4 7AL, UK
| | - David L Wild
- Warwick Systems Biology Centre and Biomedical Cell Biology, Warwick Medical School, University of Warwick, Coventry CV4 7AL, UK
| |
Collapse
|
6
|
Birlutiu A, d'Alché-Buc F, Heskes T. A Bayesian Framework for Combining Protein and Network Topology Information for Predicting Protein-Protein Interactions. IEEE/ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS 2015; 12:538-550. [PMID: 26357265 DOI: 10.1109/tcbb.2014.2359441] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.1] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/05/2023]
Abstract
Computational methods for predicting protein-protein interactions are important tools that can complement high-throughput technologies and guide biologists in designing new laboratory experiments. The proteins and the interactions between them can be described by a network which is characterized by several topological properties. Information about proteins and interactions between them, in combination with knowledge about topological properties of the network, can be used for developing computational methods that can accurately predict unknown protein-protein interactions. This paper presents a supervised learning framework based on Bayesian inference for combining two types of information: i) network topology information, and ii) information related to proteins and the interactions between them. The motivation of our model is that by combining these two types of information one can achieve a better accuracy in predicting protein-protein interactions, than by using models constructed from these two types of information independently.
Collapse
|
7
|
Verbanck M, Lê S, Pagès J. A new unsupervised gene clustering algorithm based on the integration of biological knowledge into expression data. BMC Bioinformatics 2013; 14:42. [PMID: 23387364 PMCID: PMC3635920 DOI: 10.1186/1471-2105-14-42] [Citation(s) in RCA: 18] [Impact Index Per Article: 1.6] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/04/2012] [Accepted: 01/18/2013] [Indexed: 12/03/2022] Open
Abstract
Background Gene clustering algorithms are massively used by biologists when analysing omics data. Classical gene clustering strategies are based on the use of expression data only, directly as in Heatmaps, or indirectly as in clustering based on coexpression networks for instance. However, the classical strategies may not be sufficient to bring out all potential relationships amongst genes. Results We propose a new unsupervised gene clustering algorithm based on the integration of external biological knowledge, such as Gene Ontology annotations, into expression data. We introduce a new distance between genes which consists in integrating biological knowledge into the analysis of expression data. Therefore, two genes are close if they have both similar expression profiles and similar functional profiles at once. Then a classical algorithm (e.g. K-means) is used to obtain gene clusters. In addition, we propose an automatic evaluation procedure of gene clusters. This procedure is based on two indicators which measure the global coexpression and biological homogeneity of gene clusters. They are associated with hypothesis testing which allows to complement each indicator with a p-value. Our clustering algorithm is compared to the Heatmap clustering and the clustering based on gene coexpression network, both on simulated and real data. In both cases, it outperforms the other methodologies as it provides the highest proportion of significantly coexpressed and biologically homogeneous gene clusters, which are good candidates for interpretation. Conclusion Our new clustering algorithm provides a higher proportion of good candidates for interpretation. Therefore, we expect the interpretation of these clusters to help biologists to formulate new hypothesis on the relationships amongst genes.
Collapse
Affiliation(s)
- Marie Verbanck
- Applied Mathematics Department, Agrocampus Ouest, 65, rue de Saint-Brieuc, Rennes, France.
| | | | | |
Collapse
|
8
|
Rantalainen M, Holmes CC. Accounting for control mislabeling in case-control biomarker studies. J Proteome Res 2011; 10:5562-7. [PMID: 22010953 DOI: 10.1021/pr200507b] [Citation(s) in RCA: 4] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/30/2022]
Abstract
In biomarker discovery studies, uncertainty associated with case and control labels is often overlooked. By omitting to take into account label uncertainty, model parameters and the predictive risk can become biased, sometimes severely. The most common situation is when the control set contains an unknown number of undiagnosed, or future, cases. This has a marked impact in situations where the model needs to be well-calibrated, e.g., when the prediction performance of a biomarker panel is evaluated. Failing to account for class label uncertainty may lead to underestimation of classification performance and bias in parameter estimates. This can further impact on meta-analysis for combining evidence from multiple studies. Using a simulation study, we outline how conventional statistical models can be modified to address class label uncertainty leading to well-calibrated prediction performance estimates and reduced bias in meta-analysis. We focus on the problem of mislabeled control subjects in case-control studies, i.e., when some of the control subjects are undiagnosed cases, although the procedures we report are generic. The uncertainty in control status is a particular situation common in biomarker discovery studies in the context of genomic and molecular epidemiology, where control subjects are commonly sampled from the general population with an established expected disease incidence rate.
Collapse
Affiliation(s)
- Mattias Rantalainen
- Department of Statistics, University of Oxford, 1 South Parks Road, Oxford, OX1 3TG, United Kingdom
| | | |
Collapse
|
9
|
|
10
|
Xu Q, Xiang EW, Yang Q. Transferring network topological knowledge for predicting protein-protein interactions. Proteomics 2011; 11:3818-25. [DOI: 10.1002/pmic.201100146] [Citation(s) in RCA: 5] [Impact Index Per Article: 0.4] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/21/2011] [Revised: 06/03/2011] [Accepted: 06/22/2011] [Indexed: 02/01/2023]
|
11
|
Freschi V. Improved biological network reconstruction using graph Laplacian regularization. J Comput Biol 2011; 18:987-96. [PMID: 21702693 DOI: 10.1089/cmb.2010.0232] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.2] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/13/2022] Open
Abstract
Biological networks reconstruction is a crucial step towards the functional characterization and elucidation of living cells. Computational methods for inferring the structure of these networks are of paramount importance since they provide valuable information regarding organization and behavior of the cell at a system level and also enable careful design of wet-lab experiments. Despite many recent advances, according to the scientific literature, there is room for improvements from both the efficiency and the accuracy point of view in link prediction algorithms. In this article, we propose a new method for the inference of biological networks that makes use of a notion of similarity between graph vertices within the framework of graph regularization for ranking the links to be predicted. The proposed approach results in more accurate classification rates in a wide range of experiments, while the computational complexity is reduced by two orders of magnitude with respect to many current state-of-the-art algorithms.
Collapse
Affiliation(s)
- Valerio Freschi
- DiSBeF-Department of Base Sciences and Fundamentals, University of Urbino, Urbino, Italy.
| |
Collapse
|
12
|
Xiong J, Yuan D, Fillingham JS, Garg J, Lu X, Chang Y, Liu Y, Fu C, Pearlman RE, Miao W. Gene network landscape of the ciliate Tetrahymena thermophila. PLoS One 2011; 6:e20124. [PMID: 21637855 PMCID: PMC3102692 DOI: 10.1371/journal.pone.0020124] [Citation(s) in RCA: 25] [Impact Index Per Article: 1.9] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/20/2011] [Accepted: 04/13/2011] [Indexed: 01/03/2023] Open
Abstract
Background Genome-wide expression data of gene microarrays can be used to infer gene networks. At a cellular level, a gene network provides a picture of the modules in which genes are densely connected, and of the hub genes, which are highly connected with other genes. A gene network is useful to identify the genes involved in the same pathway, in a protein complex or that are co-regulated. In this study, we used different methods to find gene networks in the ciliate Tetrahymena thermophila, and describe some important properties of this network, such as modules and hubs. Methodology/Principal Findings Using 67 single channel microarrays, we constructed the Tetrahymena gene network (TGN) using three methods: the Pearson correlation coefficient (PCC), the Spearman correlation coefficient (SCC) and the context likelihood of relatedness (CLR) algorithm. The accuracy and coverage of the three networks were evaluated using four conserved protein complexes in yeast. The CLR network with a Z-score threshold 3.49 was determined to be the most robust. The TGN was partitioned, and 55 modules were found. In addition, analysis of the arbitrarily determined 1200 hubs showed that these hubs could be sorted into six groups according to their expression profiles. We also investigated human disease orthologs in Tetrahymena that are missing in yeast and provide evidence indicating that some of these are involved in the same process in Tetrahymena as in human. Conclusions/Significance This study constructed a Tetrahymena gene network, provided new insights to the properties of this biological network, and presents an important resource to study Tetrahymena genes at the pathway level.
Collapse
Affiliation(s)
- Jie Xiong
- Key Laboratory of Aquatic Biodiversity and Conservation, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, China
- Graduate School of Chinese Academy of Sciences, Beijing, China
| | - Dongxia Yuan
- Key Laboratory of Aquatic Biodiversity and Conservation, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, China
- State Key Laboratory of Freshwater Ecology and Biotechnology, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, China
| | | | - Jyoti Garg
- Department of Biology and Center for Research in Mass Spectrometry, York University, Toronto, Ontario, Canada
| | - Xingyi Lu
- Key Laboratory of Aquatic Biodiversity and Conservation, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, China
- Graduate School of Chinese Academy of Sciences, Beijing, China
| | - Yue Chang
- Key Laboratory of Aquatic Biodiversity and Conservation, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, China
- Graduate School of Chinese Academy of Sciences, Beijing, China
| | - Yifan Liu
- Pathology Department, University of Michigan Medical School, Ann Arbor, Michigan, United States of America
| | - Chengjie Fu
- Key Laboratory of Aquatic Biodiversity and Conservation, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, China
| | - Ronald E. Pearlman
- Department of Biology and Center for Research in Mass Spectrometry, York University, Toronto, Ontario, Canada
| | - Wei Miao
- Key Laboratory of Aquatic Biodiversity and Conservation, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan, China
- * E-mail:
| |
Collapse
|