Computational platform to make deep learning analysis of human genomics data
2020
Max Delbrueck Center for Molecular Medicine, Berlin, Germany
Deep learning methods are computational methods that extract knowledge in a data-driven fashion from large datasets and are used to address a variety of biological questions. Unfortunately, most deep learning methods lack the flexibility to adapt to new data and test new hypotheses. In the present study, the researchers aimed at tackling this issue by using a more flexible approach by separating the pre-processing and dataset specification from the modelling part. The researchers generated Janggu, a library using python programming language that facilitates deep learning for genomics applications, aiming to ease data acquisition and model evaluation. Janggu includes dataset objects that manage the extraction and transformation of information from a range of commonly used file types. Further, these dataset objects are directly compatible with popular and published deep learning libraries which in turn reduces the software engineering effort. Finally, the genomics modelling results may be visualized and exported through commonly used formats. The researchers then validated the use of Janggu on datasets coming from different types of human cells. The Janggu tool is readily applicable and flexible to address a range of questions allowing users to more effectively concentrate on testing biological hypotheses.
Deep learning for genomics using Janggu
Altuna Akalin, Wolfgang Kopp
Added on: 12-20-2021
[1] https://www.nature.com/articles/s41467-020-17155-y
