Biomedical subjects

Hari Sundar

Publications and source records attributed to Hari Sundar.

5 recordsLinked to original sources

PubMed2026

LAMBDA: a prophage detection benchmark for genomic language models.

Transformer-based genomic sequence models represent an emerging frontier in computational biology. Yet, their embeddings have not yet shown the same level of predictive power as natural and protein language models, highlighting a gap between current implementations and theoretical promise. Existing benchmarks for DNA language models primarily focus on classifying regulatory elements in eukaryotic genomes, leaving open the fundamental question of whether these models learn sequence-level features across whole genomes. We introduce LAMBDA, a benchmark designed to rigorously evaluate genome language model embeddings through phage-bacteria sequence discrimination across four categories of increasing complexity: probing tasks, fine-tuning assessments, diagnostic tests, and genome-wide prophage detection. Our comprehensive analysis of current genomic language models provides insight into the importance of training data selection relative to model size, the need for domain-specific training, and the capabilities and limitations of genomic language models for detecting prophage sequences. This benchmark represents a challenging genomic annotation task in the bacterial domain and addresses a key computational problem with direct relevance to microbiology and medicine.

Prophages↗

PubMed2026

LAMBDA: A Prophage Detection Benchmark for Genomic Language Models.

Transformer-based genomic sequence models represent an emerging frontier in computational biology. Yet, their embeddings have not yet shown the same level of predictive power as natural and protein language models, indicating a gap between current implementations and theoretical promise. Existing benchmarks for DNA language models primarily focus on classifying regulatory elements in eukaryotic genomes, leaving open the fundamental question of whether these models learn sequence-level features across whole genomes. We introduce LAMBDA, a benchmark designed to rigorously evaluate genome language model embeddings through phage-bacteria sequence discrimination across four categories of increasing complexity: probing tasks, fine-tuning assessments, diagnostic tests, and genome-wide prophage detection. Our comprehensive analysis of current genomic language models provides novel insights into the importance of training data quality relative to model size, the need for domain-specific training, and the application of genomic language models for detecting prophage sequences. This benchmark represents a challenging genomic annotation task in the bacterial domain and addresses a key computational problem with direct relevance to microbiology and medicine.

DNA language model↗

PubMed2025

The impact of tokenizer selection in genomic language models.

MOTIVATION: Genomic language models have recently emerged as a new method to decode, interpret, and generate genetic sequences. Existing genomic language models have utilized various tokenization methods, including character tokenization, overlapping and nonoverlapping k-mer tokenization, and byte-pair encoding, a method widely used in natural language models. Genomic sequences differ from natural language because of their low character variability, complex and overlapping features, and inconsistent directionality. These features make subword tokenization in genomic language models significantly different from both traditional language models and protein language models. RESULTS: This study explores the impact of tokenization in genomic language models by evaluating their downstream performance on 44 classification fine-tuning tasks. We also perform a direct comparison of byte pair encoding and character tokenization in Mamba, a state-space model. Our results indicate that character tokenization outperforms subword tokenization methods on tasks that rely on nucleotide-level resolution, such as splice site prediction and promoter detection. While byte-pair tokenization had stronger performance on the SARS-CoV-2 variant classification task, we observed limited statistically significant differences between tokenization methods on the remaining downstream tasks. AVAILABILITY AND IMPLEMENTATION: Detailed results of all benchmarking experiments are available in https://github.com/leannmlindsey/DNAtokenization. Training datasets and pretrained models are available at https://huggingface.co/datasets/leannmlindsey. Datasets and processing scripts are available at doi: 10.5281/zenodo.16287401 and doi: 10.5281/zenodo.16287130.

Natural Language Processing↗

PubMed2005

Consistent estimation of cardiac motions by 4D image registration.

A 4D image registration method is proposed for consistent estimation of cardiac motion from MR image sequences. Under this 4D registration framework, all 3D cardiac images taken at different time-points are registered simultaneously, and motion estimated is enforced to be spatiotemporally smooth, thereby overcoming potential limitations of some methods that typically estimate cardiac deformation sequentially from one frame to another, instead of treating the entire set of images as a 4D volume. To facilitate our image matching process, an attribute vector is designed for each point in the image to include intensity, boundary and geometric moment invariants (GMIs). Hierarchical registration of two image sequences is achieved by using the most distinctive points for initial registration of two sequences and gradually adding less-distinctive points for refinement of registration. Experimental results on real data demonstrate good performance of the proposed method in registering cardiac images and estimating motions from cardiac image sequences.

Algorithms↗

PubMed2005

Efficient myocyte gene delivery with complete cardiac surgical isolation in situ.

BACKGROUND: Previously, we used cardiopulmonary bypass with incomplete cardiac isolation and antegrade administration of vector for global cardiac gene delivery. Here we present a translatable cardiac surgical procedure that allows for complete surgical isolation of the heart in situ with retrograde (through the coronary venous circulation) administration of both vector and endothelial permeabilizing agents to increase myocyte transduction efficiency. METHODS: In 6 adult dogs the heart was completely isolated with tourniquets placed around both vena cavae and cannulas and all pulmonary veins. On cardiopulmonary bypass, the aorta and pulmonary artery were crossclamped, and the heart was isolated. Crystalloid cardioplegia at 4 degrees C containing 10(13) particles of adenovirus encoding LacZ and 15 microg of vascular endothelial growth factor was infused retrograde into the coronary sinus and recirculated for a total of 30 minutes. The dogs were then weaned from cardiopulmonary bypass and allowed to recover. With a catheter, 3 control dogs underwent retrograde infusion of the same cocktail without cardiac isolation or cardiopulmonary bypass. RESULTS: Beta-galactosidase activities in the cardiopulmonary bypass group were several orders of magnitude higher in both the right and left ventricles when compared with those in the control group (P < .05). X-gal staining from the cardiopulmonary bypass group showed unequivocal evidence of myocyte gene expression globally in a significant proportion of cardiac myocytes. No myocyte gene expression was observed in the control group. CONCLUSION: A novel cardiac surgical technique has been developed. This approach with cardiac isolation and retrograde delivery of vector through the coronary sinus results in efficient myocyte transduction in an adult large animal in vivo.

Animals↗