Biology subjects

Morshedian, A.

Publications and source records attributed to Morshedian, A..

5 recordsLinked to original sources

bioRxiv2026

Learning Environment-Associated Marker Rankings with Attention-Based Graph Neural Networks

Genotype-by-environment interactions can complicate plant breeding because genetic effects may vary across weather regimes, locations, and years. Although genomic prediction models can use dense marker data to predict complex crop traits, they often provide limited information about how marker relevance changes across environmental conditions. This study introduces a graph-based framework for estimating environment-associated marker rankings from multi-environment crop trial data. We introduce an attention-based graph neural network that jointly represents SNP markers and weather-based environment vectors as interacting nodes and is designed to learn environment-associated marker rankings from supervised yield data. The model is trained using yield prediction as the supervised objective, and the learned marker-level attention signals are aggregated to estimate marker rankings overall and within groups of similar environments. We evaluate the framework on two genotype-by-environment datasets. In both datasets, we examine whether learned marker rankings are stable across independent training runs and determine the overlap between high-ranked markers and external GWAS references. In a maize dataset with data from 212 different site-years, we rank markers separately within each weather-defined environment cluster and define a global marker set as the markers shared across all clusters. This provides a structured way to examine how marker rankings vary across weather-defined groups and which markers remain important across environments. We further encode each weather variable with a separate LSTM combined by attention, which makes the contribution of individual weather variables readable, and compare the resulting variable ranking against an independent SHAP-based attribution. Genes located near the top-ranked markers are also examined for functional enrichment. Overall, the results suggest that graph-based, environment-conditioned marker attribution may complement genomic prediction as an exploratory tool for studying context-dependent genomic signals in multi-environment trials.

bioinformatics↗

bioRxiv2025

LSTM-Attention-Guided Graph Neural Networks for Integrated Genotype-Environment Modeling in Maize Yield Prediction

We present a deep-learning framework that combines an LSTM, a graph neural network (GNN), and transformer-style attention to model genotype-environment (G x E) effects for maize yield prediction. Weather data for a growing season is summarized using LSTM and encoded into a 21-dimensional embedding that is used as the environment node feature; 437,214 SNPs are summarized into 548 principal components that instantiate genotype nodes. Multi-head attention dynamically weights the edges during message passing. We compare three architectures: A (fully bipartite graph), B (A with intra-set top-k similarity within genotype and within environment), and C (B with a single learnable supernode readout that attends over all nodes after message passing). The joint representations feed a compact MLP for yield prediction. Using a forward-time split (2014-2021 train; 2022 test with unseen genotypes and unseen environments), performance improves monotonically from A to C: A (RMSE 2.7749, PCC 0.4115, R2 0.1693), B (2.3683, 0.6622, 0.4385), C (2.2120, 0.6945, 0.4823). Compared to A, C has a reduction in RMSE by 0.5629 ([~]20.3%) and an increase in PCC by 0.283 ([~]68.8%), indicating that global, content-adaptive aggregation promotes local G x E propagation. Performance of our approach remains consistent regardless of the number of genotypes per environment and has strong performance under variable or unbalanced genotype sampling expression across environments.

bioinformatics↗

bioRxiv2024

Conservation of cis-regulatory codes over half a billion years of evolution

The identification of homologous cell types across species represents a crucial step in understanding cell type evolution. The retina is particularly amenable to comparative analysis because the basic morphology, connectivity, and function of its six major cell classes have remained largely invariant since the earliest stages of vertebrate evolution. Here, we show that the retinas highly conserved cellular architecture is mirrored by deep conservation of the underlying cis-regulatory codes that control gene expression. We use comparative single-cell chromatin accessibility analysis of lamprey, fish, bird, and mammalian retinas-- representing over half a billion years of evolutionary divergence--to demonstrate cross-species conservation of cis-regulatory codes in all six retinal cell classes. This conservation persists despite extensive turnover of cis-regulatory regions between distant species. Conservation manifests as the clustering of multiple distinct high-affinity transcription factor (TF) binding sites toward the center of cell-class-specific open chromatin regions with little cross-species preservation of higher-order syntax. Hierarchical clustering of machine-learning models of retinal cis-regulatory codes from diverse species recovers six clusters corresponding to the six retinal cell classes. Thus, the retinas cellular Bauplan is controlled by cis-regulatory codes which predate the divergence of extant vertebrates and persist despite nearly complete enhancer turnover.

evolutionary biology↗

bioRxiv2023

Molecular Characterization of the Sea Lamprey Retina Illuminates the Evolutionary Origin of Retinal Cell Types

The lamprey, a primitive jawless vertebrate whose ancestors diverged from all other vertebrates over 500 million years ago, offers a unique window into the primordial formation of the retina. Using single-cell RNA-sequencing, we characterized retinal cell types in lamprey and compared their molecular differentiation and regulatory networks with those in mouse and other jawed vertebrates. Our analysis revealed six cell classes and 74 distinct cell types. We discovered multiple conserved cell types shared between jawless and jawed lineages, including notably rods and cones, ON and OFF bipolar cells, and starburst amacrine cells. The conservation of these cell types indicates their emergence early in vertebrate evolution, highlighting the primal designs of retinal circuits for the rod pathway, ON-OFF discrimination, and direction selectivity. In contrast to this evidence for conservation, the pathways of diversification for amacrine cells and retinal ganglion cells appear to have distinctly diverged between the two lineages. Furthermore, we inferred master regulators in specifying retinal cell classes in both lamprey and macaque and identified common regulatory elements across species, underscoring the ancestral nature of the molecular origins governing retinal cell classes. Altogether, our characterization of the lamprey retina illuminates the evolutionary origin of visual processing in the retina.

evolutionary biology↗

bioRxiv2022

The Emergence of the Metabolic Signaling of the Nucleoredoxin-like Genes during Evolution

The nucleoredoxin-like genes NXNL1 and NXNL2 were identified through the biological activity of rod-derived cone viability factors (RdCVF and RdCVF2), the alternatively spliced variants produced by intron retention, that mediate signaling between rod and cone photoreceptors by stimulating glucose uptake. These therapeutic genes for inherited retinal degenerations also produce by splicing thioredoxin-like proteins that reduce oxidized cysteines in photoreceptor proteins. The first NXNL genes date from the first animal phyla. Intron retention produces an active RdCVF protein in the tentacles of Hydra vulgaris, a species without eyes. A Scallop RdCVF protein is produced by ciliated photoreceptors of the retina and binds its receptor, BSG1. In the lamprey, a descendent of early vertebrates, RdCVF metabolic signaling between rod and cones is fully established. In the mouse, the production of BSG1 by photoreceptors is regulated by cell-specific splicing inhibition. RdCVF signaling predates photoreceptors and evolved through two alternative splicing events.

evolutionary biology↗