Biology subjects

Gomez Ortega, J.

Publications and source records attributed to Gomez Ortega, J..

4 recordsLinked to original sources

bioRxiv2026

Temporal AI model predicts drivers of cell state trajectories across human aging

Foundational AI models have recently shown promise for predicting the impact of perturbations on cell states. However, current models typically consider only one cell state at a time, limiting their ability to learn how cellular responses unfold over time, particularly across long trajectories such as diseases of aging. Here, we develop a temporal AI model, MaxToki, trained on nearly 1 trillion gene tokens including cell state trajectories across the human lifespan to generate cell states across long timelapses of human aging. MaxToki generalized to unseen trajectories through in-context learning and predicted novel age-modulating targets that were experimentally verified to influence age-related gene programs and functional decline in vivo. MaxToki represents a promising strategy for temporal modeling to accelerate the discovery of interventions for programming therapeutic cellular trajectories.

bioinformatics↗

bioRxiv2024

Quantized multi-task learning for context-specific representations of gene network dynamics

While often represented as static entities, gene networks are highly context-dependent. Here, we developed a multi-task learning strategy to yield context-specific representations of gene network dynamics. We assembled a corpus comprising [~]103 million human single-cell transcriptomes from a broad range of tissues and diseases and performed a two stage pretraining, first with non-malignant cells to generate a foundational model and then with continual learning on cancer cells to tune the model to the cancer domain. We performed multi-task learning with the foundational model to learn context-specific representations of a broad range of cell types, tissues, developmental stages, and diseases. We then leveraged the cancer-tuned model to jointly learn cell states and predict tumor-restricting factors within the colorectal tumor microenvironment. Model quantization allowed resource-efficient fine-tuning and inference while preserving biological knowledge. Overall, multi-task learning enables context-specific disease modeling that can yield contextual predictions of candidate therapeutic targets for human disease.

bioinformatics↗

bioRxiv2023

Reciprocal inflammatory signals establish profibrotic cross-feeding metabolism

Idiopathic pulmonary fibrosis (IPF) is a disease of progressive lung remodeling and collagen deposition that leads to respiratory failure. Myeloid cells are abundant in IPF lung and in murine lung fibrosis, but their functional effects are incompletely understood. Using mouse and human lung models, we show that ornithine produced by myeloid cells expressing Arg1 serves as a substrate for proline and collagen synthesis by lung fibroblasts. The predominant Arg1-expressing myeloid cells in mouse lung were macrophages, but in IPF lung, high-dimensional imaging revealed ARG1 to be expressed mainly in neutrophils. Arg1 inhibition suppressed both ornithine levels and collagen expression in cultured, precision-cut IPF lung slices and in murine lung fibrosis. These results were confirmed in macrophage-specific Arg1 KO mice. Furthermore, we find that this pathway is regulated by cell-to-cell crosstalk, starting with purinergic signaling: Fibroblast eATP receptor P2rx4 was necessary for fibroblast IL-6 expression, which in turn was necessary for Arg1 expression by myeloid cells. Taken together, our findings define an immune-mesenchymal circuit that governs profibrotic metabolism in lung fibrosis.

immunology↗

bioRxiv2021

The Biosynthetic Costs of Amino Acids at the Base of the Food Chain Determine Their Use in Higher-order Consumer Genomes

Dietary nutrient composition is essential for shaping important fitness traits and behaviours. Many organisms are protein limited and for Drosophila melanogaster, this limitation manifests at the level of the single most limiting essential Amino Acid (AA) in the diet. The identity of this AA and its effects on female fecundity is readily predictable by a procedure called exome matching in which the sum of AAs encoded by a consumers exome is used to predict the relative proportion of AAs required in its diet. However, the exome matching calculation does not weight AA contributions to the overall profile by protein size or expression. Here we update the exome matching calculation to include these weightings. Surprisingly, although nearly half of the transcriptome is differentially expressed when comparing male and female flies, we found that creating transcriptome-weighted exome matched diets for each sex did not enhance their fecundity over that supported by exome matching alone. These data indicate that while organisms may require different amounts of dietary protein across conditions, the relative proportion of the constituent AAs remains constant. Interestingly, we also found remarkable conservation of exome matched AA profiles across taxa and that the composition of these profiles could be explained by the metabolic costs of microbial AA synthesis. Thus, it appears that bioenergetic constraints amongst autotrophs shape the relative proportion of AAs that are available across trophic levels and that that this constrains biomass composition.

genomics↗