Biology subjects

Ray, M. S.

Publications and source records attributed to Ray, M. S..

3 recordsLinked to original sources

bioRxiv2026

Multimodal Fusion of Circular Functional Data on High-resolution Neuroretinal Phenotypes

Progressive optic neuropathies, particularly glaucoma, represent a significant global health challenge, and the need for precise understanding of heterogeneous neurodegenerative phenotypes cannot be overstated. Here, we brought together two complementary sources of unstructured yet clinically relevant information about neuroretinal rim (NRR) thinning, a common clinical marker of such decay. These are based on a new dataset of fundus digital images and a corresponding dataset of optical coherence tomography, both collected from a large clinical cohort of healthy eyes. First, we represented them using a common data structure that imposed a high-resolution scale of 180 equally spaced and registered measurements on a 360{degrees} circular axis. We modeled the NRR measurements of each eye as circular curves and aligned these multimodal curves to obtain a fused NRR curve for each eye. Unsupervised clustering of these fused curves identified four clusters of eyes with structural heterogeneity, which were also found to have distinctive clinical covariates. Computation of functional derivatives revealed troughs in the curves of each cluster. Using circular statistics, we estimated the directional distributions of these troughs as potentially clinically relevant regions of NRR degeneration. A comparative study using landmark registration based on functional canonical correlation analysis demonstrated that our curve-alignment-based multimodal fusion is superior. Moreover, it improves the robustness of baseline NRR data obtained from fundus imaging.

bioinformatics↗

bioRxiv2026

Predicting Clinical Phenotypes by Growth Curve Modeling of Transcriptomic Signatures during Disease Progression

High-throughput transcriptomic analysis has benefited from many statistical tests of differential gene expression across two or more groups such as t tests, ANOVA, etc. Yet, in complex transcriptomic datasets such as multi-group longitudinal measures, few studies have addressed such key issues as group effects and temporal dependency in expression profiles with a single model that is both practically effective and theoretically grounded. In this study, we used Growth Curve Model (GCM), as a generalization of MANOVA, to identify differentially expressed longitudinal profiles of genes, and thus predicted the associated clinical phenotypes, of pediatric lupus during the progressions of the disease across two different racial groups. In particular, we detected a module of histone genes which was shown to be linked with lupus.

bioinformatics↗

bioRxiv2025

Automated generation of personalized trajectories of aging phenotypes with DyViA-GAN

With a general increase in human lifespan, the need for technological advances to develop strategies for healthy aging has assumed great importance. In the present study, our goal is to predict the progression of selected aging phenotypes in a given healthy individual as one continues aging past 65 years. Therefore, we developed a novel framework called Dynamic Views of Aging with conditional Generative Adversarial Networks (or DyViA-GAN) which is capable of predicting the plausible personalized trajectories of a selected aging phenotype conditioned on the available measurements of the phenotype at a few initial time instances, and additional covariates. Given the prevalence of osteoporosis in the aging population, we selected femoral neck Bone Mineral Density (BMD) of a healthy individual as the phenotype of interest, and baseline individual Body Mass Index (BMI) as covariate. We trained DyViA-GAN on a publicly available longitudinal dataset of a large cohort of mostly white women in the United States of age 65 years or above. Thus, it generated, for each individual, continuous phenotype trajectories, along with a corresponding region of acceptable predictions, for an age range of 66 to 89 years, for eight different combinations both with and without involving the covariate. The prediction results were subjected to rigorous quality-control and multiple comparative analyses. Our results clearly demonstrate the potential of generative deep learning frameworks in healthspan research.

bioinformatics↗