Abstract
This special issue includes a selection of articles presented at the 13th International Conference on Computational Advances in Bio and Medical Sciences (ICCABS 2025), held on January 12–14, 2025, in Atlanta, Georgia, USA. The conference was hosted by the Department of Computer Science at Georgia State University, with Professors Alexander Zelikovsky from Georgia State University and Sanguthevar Rajasekaran from the University of Connecticut serving as General Chairs, and Professors Murray Patterson from Georgia State University and Shibu Yooseph from Claremont McKenna College serving as Program Committee Chairs.
ICCABS has the goal of bringing together researchers, scientists, and students from academia, laboratories, and industry to discuss recent advances in computational techniques and applications in the areas of biology, medicine, and drug discovery. In 2025, the technical program of the conference included 15 talks selected from submissions to Track 1 of the main conference. The program also included invited talks presented at four co-located workshops: eight talks presented 13th Workshop on Computational Advances for Next Generation Sequencing (CANGS 2025), 22 talks presented at the 12th Workshop on Computational Advances in Molecular Epidemiology (CAME 2025), three talks presented at the 6th Workshop on Computational Advances for Single-Cell Omics Data Analysis (CASCODA 2025), and 15 talks presented at the 1st Workshop on Metagenomics Research and Applications (CAMeRA 2025). Extended abstracts of the 15 Track 1 talks and 11 of the workshop talks appear in the ICCABS 2025 post-proceedings published as volume 15,599 of Springer Verlag’s Lecture Notes in Bioinformatics series.
This special issue includes a selection of nine articles presented at ICCABS 2025. The first five articles were presented in Track 1 of the ICCABS conference. The first article, by Rossignolo and Comin, introduces USTAR-CR, a fast and space-efficient algorithm for compressing multiple k-mer sets. Experimental results on real sequencing datasets show that USTAR-CR achieves superior compression ratios and up to 64x speedups compared to the state-of-the-art tool GGCAT. The second article, by Gemin, Pizzi, and Comin, introduces DuoHash, a framework that enables the efficient computation of hash functions for spaced seeds. Experimental results show that DuoHash substantially outperforms existing algorithms, achieving speedups of up to 11x on short reads with a spaced seed of medium density. The third article, by Altayyar and Liao, presents FDS-CAP, a novel graph-based deep learning framework for comorbidity prediction. Cross-validation experiments on benchmarking data show that FDS-CAP consistently outperforms state-of-the-art geometric embedding methods for comorbidity prediction, with an AUROC of 0.966. In the fourth article, Makohon et al. investigate the use of chain-of-thought and knowledge graph prompt engineering to improve the quality of clinical notes generated using large language models. Experimental evaluation on clinical cases from the CodiEsp benchmark shows that the proposed approach outperforms clinical notes generated by GPT-4 using standard one-shot prompts. In the fifth article, Mehta et al. present the MEditome bioinformatics pipeline for the identification of RNA edit sites in microbiomes by leveraging de novo genome assembly to reduce potential biases introduced by the use of reference genomes. The authors also present the results of validation experiments conducted on sequencing data from the Integrative Human Microbiome Project (iHMP2).
The last four articles were presented at the joint ICCABS workshops. The article by Abdelnaby and Moussa, presented at the CASCODA workshop, details the results of a comprehensive evaluation of PCA and random projection methods commonly used for single-cell RNA-Seq data analysis. The authors also introduce a novel Matching Sparsity Random Projection algorithm, shown to outperform PCA not only in running time but also in several clustering-based accuracy metrics. The article by Murad et al., presented at the CAME workshop, introduces DANCE, a framework for T cell receptor classification combining Chaos game representation of protein sequences with deep learning image classifiers. Evaluation experiments on TCR sequences from TCRdb show that DANCE outperforms in accuracy several sequence-based classifiers. The article by Adeniyi et al., presented at the CAMeRA workshop, introduces a Hidden Markov Model for modeling temporal dynamics of epistatic interactions between single amino acid variants in SARS-CoV-2. The model is validated on a large dataset of SARS-CoV-2 spike protein sequences and shown to detect Alpha variant epistatic networks months before widespread circulation. The article by Sims, Khudyakov, and Zelikovsky, also presented at the CAMeRA workshop, evaluates calibration of Illumina MiSeq base-quaity scores by leveraging overlaps in paired-end reads generated from short amplicons. The authors also introduce a maximum likelihood model for recalibration of base quality scores.
We would like to thank the Editor-in-Chief Mona Singh for providing us with the opportunity to highlight some of the exciting research presented at ICCABS 2025 in the Journal of Computational Biology. We would also like to thank the ICCABS Program Committee and the anonymous reviewers for volunteering their time to review the selected papers. Finally, we would like to thank all ICCABS authors—the conference could not continue to thrive without their high-quality contributions.