Step 3: Using Time-Series Data
Time-series data gives information about the values of genes at a series of consecutive time points. This temporal information can be exploited to infer directionality of edges, or help to infer causal relations between genes. However, adding temporal information also creates a more complex dataset. It adds interdependencies between experiments (time-points) that don’t exist in steady-state data, so more care has to be taken in analysis. Three types of algorithms will be presented in this section: mutual information, ordinary differential equations with l1 regularization, and dynamic Bayesian Networks. Each of these approaches makes different assumptions about the data.
KeywordsCovariance Shrinkage Lasso
Unable to display preview. Download preview PDF.