The advancements of high-throughput sequencing (HTS) technologies produce enormous amounts of omics data, which allows us to get new insights into genome and enhance our understanding of gene regulation and disease pathogenesis. Currently, the continuous accumulation of these data has led to two main trends in computational genomics. Firstly, there is a rapid growth in sample size and data dimensionality. The development of HTS technologies-- from bulk, single cell, all the way to the newly emerged spatial genomics and spatial transcriptomics—has led to dramatic increase in the sample size, dimensionality and sparsity of genomic and transcriptomic data. This poses great challenges for bioinformatics to reveal structure, function, and disorder of genome by mining these growing data. Secondly, multi-modal integrative analysis has become essential. Gene expressions are the results of cross talk between multiple molecular layers, the integrative analysis of multi-omics and even with biomedical data holds the promise of unravelling the complex gene regulation and enabling personalized medicine. Inevitably, this integration and analysis also faces challenges such as the heterogeneous nature of data, lack of simultaneity, and varying distributional properties across multi-modal data.
Artificial intelligence (AI) methods with Machine Learning (ML), especially Deep Learning (DL), as its core has recently revolutionized the fields ranging from computer vision to natural language processing. When applied to computational genomics, deep neural network can learn from extensive omics datasets with a great number of samples serving as training data. This encourage networks, such as gene regulatory network, to model complex patterns and find biological pathways underlying noisy and large-scale data. Besides, there is a growing consensus towards incorporating multi-omics and even biomedical data into a multi-modal network. This kind of network is expected to enhance our current knowledge of genomic structural changes, gene-environment interactions, gene regulatory mechanisms, and genetic disorders during diseases development.
Thus, the topic is interested in the advancements and applications of DL and ML methods for the following issues (not limited to):
• Calling and differential analysis of compartment, TAD, or loop structures of genome on Hi-C data.
• Analysis of structures and functions of genome. (e.g., DNA-seq, RNA-seq, Hi-C, ATAC-seq, or ChIP-seq etc.)
• Prediction of RNA secondary and tertiary structures, facilitating the understanding of RNA-RNA/RNA-protein relationship.
• Analysis of spatial transcriptomics, spatial epigenetics, or spatial proteomics. (e.g., spatial RNA-seq, spatial ATAC-seq, or spatial protein data (CODEX) etc.)
• Multi-omics data integration, interpretation, and its application.
• Analysis of biomedical data. (e.g., biomedical imaging, electronic medical records (EMRs), clinical data, or sequencing data etc.)
• Biomedical regulatory network for molecular mechanism annotation.
• Predictions of ncRNA-disease interactions by leveraging sequencing data, structural information, and functional annotations, leading to the discovery of novel therapeutic targets.
• Integrative analysis of omics and biomedical data for biomarker discovery as well as disease prognostics, diagnostics, and treatment.
• Prospects of future advancements in AI for bioinformatics.
All types of articles related to the above issues are welcomed. Data involved in articles can be at any one or more levels, such as bulk, single cell, or spatial levels. And papers must be original and not under review elsewhere.

Submit your idea

Impact

Frontiers in Genetics

Submit your idea

Impact