Other applications of deep learning in bioinformatics

There are many applications of deep learning in bioinformatics:

  • Protein structure prediction: Deep learning can be used to infer the three-dimensional structure of a protein from its sequence, which is important for understanding protein function and interactions. For example, AlphaFold1 is a deep learning-based method that achieved breakthrough results in the international protein structure prediction competition CASP.
  • Drug discovery: Deep learning can be used to design new drug molecules, predict drug activity, toxicity, metabolism and other properties, and find drug targets and ligands. For example, DeepChem2 is an open source deep learning platform that provides a series of tools and models to process chemical and biological data.
  • Gene editing: Deep learning can be used to optimize gene editing technologies, such as CRISPR/Cas9, to improve its targeting efficiency and reduce its off-target risk. For example, DeepCRISPR3 is a deep learning-based method that can predict the editing effects of the CRISPR/Cas9 system in different cell types.
  • Epigenetics: Deep learning can be used to analyze epigenetic data, such as DNA methylation, histone modifications, and chromatin accessibility, to reveal their relationship with gene expression, transcriptional regulation, and diseases. For example, DeepChrome4 is a deep learning-based method that can predict gene expression levels from histone modification data.