A special issue of Genome Research, published in collaboration with the International Conference on Research in Computational Molecular Biology (RECOMB), showcases 20 innovative computational methods with broad applications in genomics. These methods address challenges in diverse areas such as spatial transcriptomics, single-cell sequencing, long-read sequencing, genomic variation analysis, and metagenomics. Highlighted studies in the issue introduce novel algorithms and techniques aimed at improving privacy, accuracy, and scalability in genomic research.
Predicting Disease Risk Without Longitudinal Data
One of the featured methods, PRiMeR, developed by Sens et al. (2024), focuses on predicting disease risk by utilizing genetic data in a novel way. PRiMeR avoids the need for traditional longitudinal studies by training on genetic data from healthy cohorts, along with genome-wide association studies (GWAS) results, to predict disease risk for new patients. This approach was successfully validated using simulations of type 2 diabetes and neurodegenerative diseases, including Alzheimer’s and Parkinson’s. PRiMeR provides a faster and potentially more effective route for identifying at-risk individuals, which may enable the development of timely prevention strategies.
Federated Algorithms for Privacy-Preserving Genomic Analysis
In a different study, Hong et al. (2024) developed SF-Relate, a secure federated algorithm designed to identify genetic relatives across distributed genomic datasets while preserving individual privacy. SF-Relate employs a combination of hashing and bucketing strategies, which differentiate relatives from nonrelatives without directly accessing sensitive genomic data. The method provides a way to exclude close relatives from genomic studies—a necessary step to avoid introducing bias—while maintaining robust privacy protections for participants. The ability to securely estimate kinship could be crucial in future large-scale genomic studies.
Understanding the Role of Circular Extrachromosomal DNA in Cancer
Extrachromosomal DNA (ecDNA) has emerged as a significant player in oncogene amplification across various cancers. Two new methods, CoRAL (Zue et al. 2024) and Decoil (Giurgiu et al. 2024), employ long-read sequencing to analyze the structure and dynamics of ecDNA in tumors. These methods provide insights into how ecDNA influences tumor growth, evolution, and resistance to treatment.
Single-Cell Methods for Understanding Cell-Cell Interactions
DIISCO (Park et al. 2024), another study featured in the special issue, addresses the need for analyzing temporal dynamics of cell-cell interactions in complex biological systems. Using single-cell RNA sequencing data, DIISCO reveals how cells interact and communicate during normal and disease processes. When tested on simulated and experimental lymphoma–immune interaction data, the method identified immune interactions involving a cytotoxic T cell subtype that expands in response to therapy. This knowledge could be instrumental in guiding the development of improved immunotherapies that target specific cell interactions.
Advancing Spatial Transcriptomics
The issue also highlights SpaCeNet (Schrod et al. 2024), a tool designed to analyze spatial transcriptomics data and reconstruct networks of both intracellular and intercellular interactions at single-cell resolution. SpaCeNet was applied to datasets from mouse visual cortex, mouse organoids, and Drosophila blastoderm, revealing complex spatial organization patterns in cell populations. Understanding these patterns can allow researchers to gain new perspectives on processes related to cellular growth and development, as well as disease progression.
Overcoming Challenges in Metagenomic Data Analysis
Metagenomics, the study of genetic material from environmental samples, faces particular difficulties due to repetitive DNA sequences that complicate genome assembly. GraSSRep (Azizpour et al. 2024), introduced in this issue, provides a novel method for detecting and classifying repetitive DNA in metagenomic datasets. This innovation is particularly relevant in microbial community studies, where genome dynamics such as horizontal gene transfer and gene duplication further complicate accurate sequence assembly.
Innovations Across Genomics Fields
In addition to these highlighted studies, the special issue of Genome Research features several other computational methods that advance research in cancer genomics, transcriptomics, gene regulatory networks, genomic variation, and biomolecular representation learning. These methods collectively push the boundaries of how genomic data is analyzed and interpreted, providing scientists with new tools to explore the complexities of biological systems.
Predicting Disease Risk Without Longitudinal Data
One of the featured methods, PRiMeR, developed by Sens et al. (2024), focuses on predicting disease risk by utilizing genetic data in a novel way. PRiMeR avoids the need for traditional longitudinal studies by training on genetic data from healthy cohorts, along with genome-wide association studies (GWAS) results, to predict disease risk for new patients. This approach was successfully validated using simulations of type 2 diabetes and neurodegenerative diseases, including Alzheimer’s and Parkinson’s. PRiMeR provides a faster and potentially more effective route for identifying at-risk individuals, which may enable the development of timely prevention strategies.
Federated Algorithms for Privacy-Preserving Genomic Analysis
In a different study, Hong et al. (2024) developed SF-Relate, a secure federated algorithm designed to identify genetic relatives across distributed genomic datasets while preserving individual privacy. SF-Relate employs a combination of hashing and bucketing strategies, which differentiate relatives from nonrelatives without directly accessing sensitive genomic data. The method provides a way to exclude close relatives from genomic studies—a necessary step to avoid introducing bias—while maintaining robust privacy protections for participants. The ability to securely estimate kinship could be crucial in future large-scale genomic studies.
Understanding the Role of Circular Extrachromosomal DNA in Cancer
Extrachromosomal DNA (ecDNA) has emerged as a significant player in oncogene amplification across various cancers. Two new methods, CoRAL (Zue et al. 2024) and Decoil (Giurgiu et al. 2024), employ long-read sequencing to analyze the structure and dynamics of ecDNA in tumors. These methods provide insights into how ecDNA influences tumor growth, evolution, and resistance to treatment.
Single-Cell Methods for Understanding Cell-Cell Interactions
DIISCO (Park et al. 2024), another study featured in the special issue, addresses the need for analyzing temporal dynamics of cell-cell interactions in complex biological systems. Using single-cell RNA sequencing data, DIISCO reveals how cells interact and communicate during normal and disease processes. When tested on simulated and experimental lymphoma–immune interaction data, the method identified immune interactions involving a cytotoxic T cell subtype that expands in response to therapy. This knowledge could be instrumental in guiding the development of improved immunotherapies that target specific cell interactions.
Advancing Spatial Transcriptomics
The issue also highlights SpaCeNet (Schrod et al. 2024), a tool designed to analyze spatial transcriptomics data and reconstruct networks of both intracellular and intercellular interactions at single-cell resolution. SpaCeNet was applied to datasets from mouse visual cortex, mouse organoids, and Drosophila blastoderm, revealing complex spatial organization patterns in cell populations. Understanding these patterns can allow researchers to gain new perspectives on processes related to cellular growth and development, as well as disease progression.
Overcoming Challenges in Metagenomic Data Analysis
Metagenomics, the study of genetic material from environmental samples, faces particular difficulties due to repetitive DNA sequences that complicate genome assembly. GraSSRep (Azizpour et al. 2024), introduced in this issue, provides a novel method for detecting and classifying repetitive DNA in metagenomic datasets. This innovation is particularly relevant in microbial community studies, where genome dynamics such as horizontal gene transfer and gene duplication further complicate accurate sequence assembly.
Innovations Across Genomics Fields
In addition to these highlighted studies, the special issue of Genome Research features several other computational methods that advance research in cancer genomics, transcriptomics, gene regulatory networks, genomic variation, and biomolecular representation learning. These methods collectively push the boundaries of how genomic data is analyzed and interpreted, providing scientists with new tools to explore the complexities of biological systems.