Method for analysis of omics data

A method for analysis of omics data related to samples derived from a human, animal or disease model, comprising:

• a) Determining a plurality of latent spaces by inputting the omics data into a plurality of omics neural networks in parallel and achieving a plurality of latent spaces each comprising a set of latent factors by performing the calculations of the plurality of omics neural networks in a data processing unit, wherein the output of the omics neural networks each provides a respective latent space;
• b) Consolidating the plurality of latent spaces into one reproducible latent space, said consolidating including:
• selecting latent factors which are similarly present in a minimum number of the plurality of latent spaces by comparing each latent factor of each of the latent spaces with each latent factor from each of the other latent spaces using a similarity or distance metric and selecting the latent factors which are present in at least a defined number N (N=integer number) of latent spaces within a given similarity range;
• clustering the selected latent factors by grouping of latent factors which are present in at least a defined number N of the latent spaces together into clusters such that each cluster contains those latent factors from many latent spaces which are similar to each other according to said similarity or distance metric and thresholds;
• and
• calculating a respective average of the selected latent factors in a cluster for each of the clusters wherein the resulting average is a single latent factor related to a respective cluster, and wherein the set of single latent factors for the plurality of clusters form a consolidated latent space of the plurality of omics neural networks,
• wherein the step of consolidating is processed by a data processing unit;
• wherein the consolidated latent factors include gene expressions and represent characteristics of the samples to be analyzed as biomarkers measured for the sample as a result of the consolidated latent factors,
• c) processing the stored consolidated latent factors forming a consolidated latent space obtained from the plurality of omics neural networks to at least one of i) detect disease processes, ii) predict drug responses, iii) predict patient survival probabilities or iv) find similar patients in a cohort.