Fig. 1

Tutorial for genomic data integration. The tutorial of best practices presents the different steps to conduct multi-omics integrations: A, step 1 - Constructing the genomic data matrix. B, step 2 - Defining a clear and precise question of interest where biological questions concern describing omics interactions and interplay, selecting biomarkers specific to a trait, or predicting phenotypes from omics; C, step 3 - Selecting the tool, by considering tools’ specificities such as their coding language, the accessibility or not to their source code, the quality of their documentation and frequency of their updates, their methods’ main concepts and data requirements (see Table 1); D, step 4 - Preprocessing data, especially to remove or impute missing values, identify then remove outliers or reduce their impact, normalize data, correct batch effects; E, step 5 - Pre-analyzing data, by first importing data in the expected format with the right dimensions and types, then analyzing them by variable (univariate analysis) and dataset (multivariate analysis) to reveal major insights. F, step 6 - Genomic data integration for data description, selection and prediction.