Deep learning for genomics using Janggu

Abstract

In recent years, numerous applications have demonstrated the potential of deep learning for an improved understanding of biological processes. However, most deep learning tools developed so far are designed to address a specific question on a fixed dataset and/or by a fixed model architecture. Here we present Janggu, a python library facilitates deep learning for genomics applications, aiming to ease data acquisition and model evaluation. Among its key features are special dataset objects, which form a unified and flexible data acquisition and pre-processing framework for genomics data that enables streamlining of future research applications through reusable components. Through a numpy-like interface, these dataset objects are directly compatible with popular deep learning libraries, including keras or pytorch. Janggu offers the possibility to visualize predictions as genomic tracks or by exporting them to the bigWig format as well as utilities for keras-based models. We illustrate the functionality of Janggu on several deep learning genomics applications. First, we evaluate different model topologies for the task of predicting binding sites for the transcription factor JunD. Second, we demonstrate the framework on published models for predicting chromatin effects. Third, we show that promoter usage measured by CAGE can be predicted using DNase hypersensitivity, histone modifications and DNA sequence features. We improve the performance of these models due to a novel feature in Janggu that allows us to include highorder sequence features. We believe that Janggu will help to significantly reduce repetitive programming overhead for deep learning applications in genomics, and will enable computational biologists to rapidly assess biological hypotheses.

ReFeaFi: Genome-wide prediction of regulatory elements driving transcription initiation

“…To further improve the results, it is necessary to consider the cell type when predicting promoters and enhancers. It has been shown that promoter prediction can also be improved using extra information as input, for example chromatin data 39 or possibly applying nonparametric methods as described and tested on promoter regions of a model dicot plant Arabidopsis thaliana 40 . However, the approach described in this paper has the advantage that it is very general.…”

Deep learning in next-generation sequencing

“…With the growing popularity of DL methods for analyzing sequencing data, several software frameworks and packages specifically designed for bioinformatics data have been introduced recently. Libraries, such as Nucleus [48] or Janggu [49] , can be used alongside Keras, TF, or PyTorch. They offer dedicated objects for processing biological sequence data, which makes it easy to read, write, analyze, and visualize data in common genomics file formats, such as BAM, FASTA, bigWig, VCF, or BED.…”

immuneML: an ecosystem for machine learning analysis of adaptive immune receptor repertoires

“…Tailored ML frameworks and platforms that account for the idiosyncrasies of the underlying data have been published for applications in genomics 38,39 , proteomics 40,41 , biomedicine 42 , and chemistry 43 . Their creation recognizes the infeasibility to define, implement, and train appropriate ML models by relying solely on generic ML frameworks such as scikit-learn 44 or PyTorch 45 .…”