Using deep neural networks trained on massive multi-modal inputs for ageing research
Quentin Vanhaelen
Insilico Medicine inc.
Aging is a complex process and occurs at different rates and to different extents in various organ systems. One problem is that the evaluation of aging changes and possible anti-aging therapies requires a comprehensive set of robust biomarkers. However, most of these biomarkers are not representative of the health state of the entire organism or individual systems and are not easily measured or targeted with known interventions. Deep learning technique based on a modular ensemble of multiple Deep Neural Networks (DNNs) stacked into an ensemble and trained on tens of thousands of blood biochemistry samples can be used for predicting human chronological age. The analysis of relative feature importance within the DNNs helps to deduce the most important features that may shed light on the contribution of these systems to the aging process. Albumin, glucose, alkaline phosphatase, urea and erythrocytes were identified as being the five most markers for predicting human chronological age.
Disclosure: The author is an employee of InSilico Medicine inc.
Quentin Vanhaelen
Insilico Medicine inc.
Aging is a complex process and occurs at different rates and to different extents in various organ systems. One problem is that the evaluation of aging changes and possible anti-aging therapies requires a comprehensive set of robust biomarkers. However, most of these biomarkers are not representative of the health state of the entire organism or individual systems and are not easily measured or targeted with known interventions. Deep learning technique based on a modular ensemble of multiple Deep Neural Networks (DNNs) stacked into an ensemble and trained on tens of thousands of blood biochemistry samples can be used for predicting human chronological age. The analysis of relative feature importance within the DNNs helps to deduce the most important features that may shed light on the contribution of these systems to the aging process. Albumin, glucose, alkaline phosphatase, urea and erythrocytes were identified as being the five most markers for predicting human chronological age.
Disclosure: The author is an employee of InSilico Medicine inc.
