Deep Integrated Biomarkers of Aging
- 804 Downloads
Recent advances in deep learning (DL) and other machine learning methods (ML) and the availability of large population data sets annotated with age and other features led to the development of the many predictors of chronological age frequently referred to as the deep aging clocks (DACs). Many of these aging clocks were pioneered by Insilico Medicine and are available for testing online via www.young.ai, www.aging.ai, and other resources. Several of the DACs demonstrate biological relevance and can be used for a variety of applications ranging from aging and disease target identification, identifying the age-related features implicated in diseases, the inference of causal graphs, analyze the population-specificity of the data type and analyze the mechanism of drug response and many others. The generative adversarial networks (GANs) can be used to synthesize new data of patients of specified ages in large volume providing for novel target identification techniques, as well as for anonymization and patient privacy methods. In this chapter we provide a brief overview of the DACs, their advantages, and disadvantages as well as the commentary on the challenges in this emerging field.
The authors would like to thank Nvidia Corporation and personally Mark Berger for providing early access to the graphics processing units (GPUs) that helped power the development of the DACs.
Conflict of Interest
AZ and PM work for Insilico Medicine, a for-profit longevity biotechnology company developing the end-to-end target identification and drug discovery pipeline for a broad spectrum of age-related diseases. The company applied for multiple patents covering the various methods for development and applications of the aging clocks. The company may have commercial interests in this publication.
- Artemov, A. V et al. (2016) Integrated deep learned transcriptomic and structure-based predictor of clinical trials outcomes. bioRxiv https://doi.org/10.1101/095653
- Galkin, F. et al. (2018) Human microbiome aging clocks based on deep learning and tandem of permutation feature importance and accumulated local effects. bioRxiv https://doi.org/10.1101/507780
- Kadurin A et al (2017a) The cornucopia of meaningful leads: Applying deep adversarial autoencoders for new molecule development in oncology. Oncotarget 8:10883–10890Google Scholar
- Mamoshina P et al (2018b) Blood Biochemistry Analysis to Detect Smoking Status and Quantify Accelerated Aging in Smokers. Sci. Rep, Accepted proofGoogle Scholar
- Mamoshina P et al (2018c) Machine Learning on Human Muscle Transcriptomic Data for Biomarker Discovery and Tissue-Specific Drug Target Identification. Front. Genet. 9:242Google Scholar
- Putin E et al (2016) Deep biomarkers of human aging: Application of deep neural networks to biomarker development. 8:1–13Google Scholar
- Song, J. et al. (2017) Dual Conditional GANs for Face Aging and RejuvenationGoogle Scholar