Time-lagged Ordered Lasso for network inference

Abstract

BackgroundAccurate gene regulatory networks can be used to explain the emergence of different phenotypes, disease mechanisms, and other biological functions. Many methods have been proposed to infer networks from gene expression data but have been hampered by problems such as low sample size, inaccurate constraints, and incomplete characterizations of regulatory dynamics. Since expression regulation is dynamic, time-course data can be used to infer causality, but these datasets tend to be short or sparsely sampled. In addition, temporal methods typically assume that the expression of a gene at a time point depends on the expression of other genes at only the immediately preceding time point, while other methods include additional time points without any constraints to account for their temporal distance. These limitations can contribute to inaccurate networks with many missing and anomalous links.ResultsWe adapted the time-lagged Ordered Lasso, a regularized regression method with temporal monotonicity constraints, for de novo reconstruction. We also developed a semi-supervised method that embeds prior network information into the Ordered Lasso to discover novel regulatory dependencies in existing pathways. R code is available at https://github.com/pn51/laggedOrderedLassoNetwork.ConclusionsWe evaluated these approaches on simulated data for a repressilator, time-course data from past DREAM challenges, and a HeLa cell cycle dataset to show that they can produce accurate networks subject to the dynamics and assumptions of the time-lagged Ordered Lasso regression.Electronic supplementary materialThe online version of this article (10.1186/s12859-018-2558-7) contains supplementary material, which is available to authorized users.

Gene regulatory network inference in long-lived C. elegans reveals modular properties that are predictive of novel aging genes

“…Benchmark studies suggest that inferred GRNs improve by combining results from several algorithms (Marbach et al, 2012;Hill et al, 2016) and by adding existing knowledge of regulatory interactions as a prior in the inference process (Marbach et al, 2012;Siahpirani and Roy, 2016). Therefore, we considered three NI tools that can incorporate priors in the inference process (STAR Methods): Inferelator (Arrieta-Ortiz et al, 2015), MERLIN-P (Siahpirani and Roy, 2016) and Time-lagged Ordered Lasso (Nguyen and Braun, 2018) (TOL). Unlike yeast (Costanzo et al, 2016;Kuzmin et al, 2018), C. elegans lacks a comprehensive contextual database of regulatory interactions for adult animals to be used as priors.…”

Network inference with Granger causality ensembles on single-cell transcriptomics

“…Other elements of the GLG regression framework can be adapted as well. These include placing monotonicity constraints on the temporal lag coefficients (Nguyen and Braun, 2018), adapting the kernel so that more recent samples are assigned higher weights, or implementing kernel-based generalizations of group lasso (Yuan and Lin, 2006;Lozano et al, 2009) to regularize all coefficients from individual regulators as a group instead of independent variables. The kernel-based approach at the core of SINGE provides great flexibility to adapt it to emphasize different aspects of dynamic biological processes.…”

Integrative Modelling of Gene Expression and Digital Phenotypes to Describe Senescence in Wheat

“…In order to avoid the limitations of short or sparsely distributed temporal gene expression data, we used the time-lagged ordered lasso, a regularized regression method with temporal monotonicity constraints, for de novo reconstruction of GRNs [51,52]. The time-lagged ordered lasso is based on the ordinary lasso [53], which performs feature selection and regularization for model fitting and uses an order constraint, in this case, to reflect the relative importance of the features.…”