Table 1 Comparison of the dynamic modeling approaches for temporal omics data, the detailed methods, mining tasks, and type of problems, examples, and related references
From: Computational dynamic approaches for temporal omics data with applications to systems medicine
General approaches | Examples | Type of problems, tasks | Important features and functions | Some Reference |
|---|---|---|---|---|
Math based Deterministic, static Stochastic, dynamic | Differential equations, Fourier transform, topology based matrix factorization Stochastic differential equations, Gaussian graphical models, Probabilistic Boolean networks State space model and or hidden Markov model, Markov random fields | Parameter/rate estimations, network inference, prediction, time course (I-III) Dynamic parameter estimations transition process Causal or non-causal temporal relationships | Fixed, stable parameter, structure estimation, time invaried, non-causal Direct relationship, Nonlinear or linear. Probabilistic time varied, Nonlinear or linear Direct or indirect relationship time course (I-III) | [23,24,25,26,27,28,29,30] [36,37,38,39,40] [31,32,33,34,35] [41, 42] [43] |
Statistical based Frequentist/classical Bayesian methods | Regression vector autoregressive (VAR) models, Curve fitting, spline methods, Granger causality Bayesian models (linear or nonlinear model), growth model | Parameter estimations, predictions, hypothesis testing, biomarker/target identifications Heterogeneity discovery | Explanatory relationship without prior knowledge, pure data based time course (I-III) or phenotype dependent (IV) With prior or empirical Knowledge, probabilistic | [35, 44, 45] [46] [41, 42, 47,48,49,50,51,52,53,54,55,56,57] |
Computer sciences based Machine learning, data mining discriminative generative Neural network | Unsupervised: Distance or correlation based Supervised classification with wrapper Feedback Forward NN, time recurrent NN, convolution NN, Bayesian NN | Subtypes, modular, and heterogeneity discovery, Pattern discovery and identification Dynamic changes and trajectories Complex relationship, structure | Time course (I-III) or phenotype dependent (IV) Without knowing the outcome, classes, Defined outcomes/classes conditional joint analysis Time varied or invariaed Nonlinear or linear Direct or indirect relationship, Explanatory or predictive time course (I-III) or phenotype dependent (IV) | [26,27,28,29] [58,59,60,61,62,63,64,65,66] [68,69,70,71,72,73,74,75] |
Interactions and network, pathway function based | Predictions, integrated with public databases | phenotype dependent (IV), Graphic based Causal hypothesis | Direct or indirect relationship, Nonlinear or linear integrated with public databases interactive through manually or automate | [89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109, 116, 162] [83,84,85] [86] [87] |