Biology subjects

Magzoub, M. M.

Publications and source records attributed to Magzoub, M. M..

3 recordsLinked to original sources

bioRxiv2025

Predicting dynamic expression patterns in budding yeast with a fungal DNA language model

Predicting gene expression from DNA sequence remains challenging due to complex regulatory codes. We introduce a masked DNA language model pretrained on 165 fungal genomes closely related to budding yeast that captures conserved regulatory grammar. Fine-tuning the LM on yeast RNA-seq data--including high-resolution transcriptional regulator induction time courses generated in this study--yielded Shorkie, a model that substantially improves gene expression prediction compared to baselines trained without self-supervision. Shorkie identified canonical transcription factor (TF) binding motifs and tracked their usage across induction experiments. Furthermore, Shorkie accurately predicted variant effects, outperforming leading sequence-to-expression models in cis-eQTL classification and achieving high concordance with massively parallel reporter assays. Interpretability analyses revealed Shorkies ability to resolve promoter dynamics, splicing signals, and temporal changes in regulatory motif usage. This framework demonstrates that evolutionary-scale pretraining combined with transfer learning substantially improves our ability to decode gene regulation from sequence, providing insights into noncoding variants and regulatory networks.

genomics↗

bioRxiv2023

The Molecular Architecture of Variable Lifespan in Diversity Outbred Mice

To unravel the causes and effects of aging we can monitor the time-evolution of the aging process and learn how it is structured by genetic and environmental variation before ultimately testing theories about the causal drivers of aging. Diverse Outbred (DO) mice provide widespread, yet controlled, genetic variation generating considerable variation in mouse lifespan - here, we explore the relationship between DO mouse aging and lifespan. We profiled the plasma multiome of 110 DO mice at three ages using liquid chromatography - mass spectrometry (LC-MS)-based metabolomics and lipidomics and proteomics. Individual mice varied more than two-fold in natural lifespan. The combination of known age and resulting lifespan allows us to evaluate alternative models of how molecules were related to chronological age and lifespan. The majority of the aging multiome shifts with chronological age highlighting the accelerating chemical stress of aging. In contrast, proteomic pathways encompassing both well-appreciated aspects of aging biology, such as dysregulation of proteostasis and inflammation, as well as lesser appreciated changes such as through toll-like receptor signaling, shift primarily with fraction of life lived (the ratio of chronological age to lifespan). This measure, which approximates biological age, varies greatly across DO mice creating a global disconnect between chronological and biological age. By sampling mice near their natural death we were able to detect loss-of-homeostasis signatures involving focal dysregulation of proteolysis and the secreted phosphoproteome which may be points-of-failure in DO aging. These events are succeeded by massive changes in the multiome in mices final three weeks as widespread cell death reshapes the plasma of near-death mice.

systems biology↗

bioRxiv2018

Proteogenomic view of cancer epigenetics: the impact of DNA methylation on the cancer proteome

Aberrant DNA methylation disrupts normal gene expression in cancer and broadly contributes to oncogenesis. We previously developed MethylMix, a model-based algorithmic approach to identify epigenetically regulated driver genes. MethylMix identifies genes where methylation likely executes a functional role by using transcriptomic data to select only methylation events that can be linked to changes in gene expression. However, given that proteins more closely link genotype to phenotype recent high-throughput proteomic data provides an opportunity to more accurately identify functionally relevant abnormal methylation events. Here we present ProteoMix, which refines nominations for epigenetic driver genes by leveraging quantitative high-throughput proteomic data to select only genes where DNA methylation is predictive of protein abundance. Applying our algorithm across three cancer cohorts we find that ProteoMix narrows candidate nominations, where the effect of DNA methylation is often buffered at the protein level. Next, we find that ProteoMix genes are enriched for biological processes involved in cancer including functions involved in epithelial and mesenchymal transition. ProteoMix results are also enriched for tumor markers which are predictive of clinical features like tumor stage and we find clustering on ProteoMix genes captures cancer subtypes.

bioinformatics↗