Modeling transcriptomic age using knowledge-primed artificial neural networks

Holzscheck, Nicholas; Falckenhayn, Cassandra; Söhle, Jörn; Kristof, Boris; Siegner, Ralf; Werner, André; Schössow, Janka; Jürgens, Clemens; Völzke, Henry; Wenck, Horst; Winnefeld, Marc; Grönniger, Elke; Kaderali, Lars

doi:10.1038/s41514-021-00068-5

Cited by 40 publications

(31 citation statements)

References 86 publications

(91 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Deep learning models have been successfully applied to transcriptomic and clinical blood biomarker data for age prediction 14 , 15 . For DNA methylation data, Galkin et al recently showed that a deep neural network model, DeepMAge 16 , gave slightly better prediction performance than Horvath’s model in blood samples.…”

Section: Introductionmentioning

confidence: 99%

A pan-tissue DNA-methylation epigenetic clock based on deep learning

2022

View full text Add to dashboard Cite

Several age predictors based on DNA methylation, dubbed epigenetic clocks, have been created in recent years, with the vast majority based on regularized linear regression. This study explores the improvement in the performance and interpretation of epigenetic clocks using deep learning. First, we gathered 142 publicly available data sets from several human tissues to develop AltumAge, a neural network framework that is a highly accurate and precise age predictor. Compared to ElasticNet, AltumAge performs better for within-data set and cross-data set age prediction, being particularly more generalizable in older ages and new tissue types. We then used deep learning interpretation methods to learn which methylation sites contributed to the final model predictions. We observe that while most important CpG sites are linearly related to age, some highly-interacting CpG sites can influence the relevance of such relationships. Using chromatin annotations, we show that the CpG sites with the highest contribution to the model predictions were related to gene regulatory regions in the genome, including proximity to CTCF binding sites. We also found age-related KEGG pathways for genes containing these CpG sites. Lastly, we performed downstream analyses of AltumAge to explore its applicability and compare its age acceleration with Horvath’s 2013 model. We show that our neural network approach predicts higher age acceleration for tumors, for cells that exhibit age-related changes in vitro, such as immune and mitochondrial dysfunction, and for samples from patients with multiple sclerosis, type 2 diabetes, and HIV, among other conditions. Altogether, our neural network approach provides significant improvement and flexibility compared to current epigenetic clocks for both performance and model interpretability.

show abstract

Section: Introductionmentioning

confidence: 99%

A pan-tissue DNA-methylation epigenetic clock based on deep learning

2022

View full text Add to dashboard Cite

show abstract

“…So far, molecular aging clocks have largely relied on datasets built using bulk tissue input or purified cell populations [9][10][11][12][13][27][28][29][30][31][32][33][34] . Bulk tissue profiles (and even purified populations) average the molecular profiles from many cells, integrating tissue composition changes and celltype-specific responses.…”

mentioning

confidence: 99%

Cell-type-specific aging clocks to quantify aging and rejuvenation in neurogenic regions of the brain

et al. 2022

View full text Add to dashboard Cite

The diversity of cell types is a challenge for quantifying aging and its reversal. Here we develop ‘aging clocks’ based on single-cell transcriptomics to characterize cell-type-specific aging and rejuvenation. We generated single-cell transcriptomes from the subventricular zone neurogenic region of 28 mice, tiling ages from young to old. We trained single-cell-based regression models to predict chronological age and biological age (neural stem cell proliferation capacity). These aging clocks are generalizable to independent cohorts of mice, other regions of the brains, and other species. To determine if these aging clocks could quantify transcriptomic rejuvenation, we generated single-cell transcriptomic datasets of neurogenic regions for two interventions—heterochronic parabiosis and exercise. Aging clocks revealed that heterochronic parabiosis and exercise reverse transcriptomic aging in neurogenic regions, but in different ways. This study represents the first development of high-resolution aging clocks from single-cell transcriptomic data and demonstrates their application to quantify transcriptomic rejuvenation.

show abstract

“…In order to present a complete study, a computational regression model whose task was to predict the age of the enamel based on information about its broadly understood structure was built. The proposed model is based on artificial neural networks (ANNs), which are computational models regarded as intelligent systems used to solve complex problems with nonlinear relationships using many independent parameters [ 52 , 62 , 63 , 64 ]. MLP network with 13 inputs, one hidden layer and one output ( Figure 11 ) was used to predict the day of postnatal life based on collected data.…”

Section: Resultsmentioning

confidence: 99%

Surface and Structural Studies of Age-Related Changes in Dental Enamel: An Animal Model

et al. 2022

View full text Add to dashboard Cite

In the animal kingdom, continuously erupting incisors provided an attractive model for studying the enamel matrix and mineral composition of teeth during development. Enamel, the hardest mineral tissue in the vertebrates, is a tissue sensitive to external conditions, reflecting various disturbances in its structure. The developing dental enamel was monitored in a series of incisor samples extending the first four weeks of postnatal life in the spiny mouse. The age-dependent changes in enamel surface morphology in the micrometre and nanometre-scale and a qualitative assessment of its mechanical features were examined by applying scanning electron microscopy (SEM) and atomic force microscopy (AFM). At the same time, structural studies using XRD and vibrational spectroscopy made it possible to assess crystallinity and carbonate content in enamel mineral composition. Finally, a model for predicting the maturation based on chemical composition and structural factors was constructed using artificial neural networks (ANNs). The research presented here can extend the existing knowledge by proposing a pattern of enamel development that could be used as a comparative material in environmental, nutritional, and pharmaceutical research.

show abstract

Modeling transcriptomic age using knowledge-primed artificial neural networks

Cited by 40 publications

References 86 publications

A pan-tissue DNA-methylation epigenetic clock based on deep learning

A pan-tissue DNA-methylation epigenetic clock based on deep learning

Cell-type-specific aging clocks to quantify aging and rejuvenation in neurogenic regions of the brain

Surface and Structural Studies of Age-Related Changes in Dental Enamel: An Animal Model

Contact Info

Product

Resources

About