A5044667197- Epigenetics and DNA Methylation
- Genomics and Chromatin Dynamics
- Alzheimer's disease research and treatments
- Natural Language Processing Techniques
- Neonatal Respiratory Health Research
- Neuroinflammation and Neurodegeneration Mechanisms
- Tryptophan and brain disorders
Hebrew University of Jerusalem
2020-2024
Chronological age prediction from DNA methylation sheds light on human aging, health, and lifespan. Current clocks are mostly based linear models rely upon hundreds of sites across the genome. Here, we present GP-age, an epigenetic non-linear cohort-based clock for blood, 11,910 methylomes. Using 30 CpG alone, GP-age outperforms state-of-the-art models, with a median accuracy ∼2 years held-out blood samples, both array sequencing-based data. We show that aging-related changes occur at...
The incorporation of Denoising Diffusion Models (DDMs) in the Text-to-Speech (TTS) domain is rising, providing great value synthesizing high quality speech. Although they exhibit impressive audio quality, extent their semantic capabilities unknown, and controlling synthesized speech's vocal properties remains a challenge. Inspired by recent advances image synthesis, we explore latent space frozen TTS models, which composed bottleneck activations DDM's denoiser. We identify that this contains...
Age-dependent changes in DNA methylation allow chronological and biological age inference, but the underlying mechanisms remain unclear. Using ultra-deep sequencing of >300 blood samples from healthy individuals, we show that age-dependent are regional occur at multiple adjacent CpG sites, either stochastically or a coordinated block-like manner. Deep learning analysis single-molecule patterns two genomic loci achieved accurate prediction with median error 1.46-1.7 years on held-out human...
Summary Chronological age prediction from DNA methylation sheds light on human aging, indicates poor health and predicts lifespan. Current clocks are mostly based linear models hundreds of sites, not suitable for sequencing-based data. We present GP-age, an epigenetic clock blood, that uses a non-linear cohort-based model 11,910 blood methylomes. Using 30 CpG sites alone, GP-age outperforms state-of-the-art models, with median accuracy ~2 years held-out samples, both array show aging-related...