Extracting biological age from biomedical data via deep learning: too much of a good thing?

Age-related physiological changes in humans are linearly associated with age. Naturally, linear combinations of physiological measures trained to estimate chronological age have recently emerged as a practical way to quantify aging in the form of biological age. In this work, we used one-week long physical activity records from a 2003-2006 National Health and Nutrition Examination Survey (NHANES) to compare three increasingly accurate biological age models: the unsupervised Principal Components Analysis (PCA) score, a multivariate linear regression, and a state-of-the-art deep convolutional neural network (CNN). We found that the supervised approaches produce better chronological age estimations at the expense of a loss of the association between the aging acceleration and all-cause mortality. Consequently, we turned to the NHANES death register directly and introduced a novel way to train parametric proportional hazards models suitable for out-of-the-box implementation with any modern machine learning software. As a demonstration, we produced a separate deep CNN for mortality risks prediction that outperformed any of the biological age or a simple linear proportional hazards model. Altogether, our findings demonstrate the emerging potential of combined wearable sensors and deep learning technologies for applications involving continuous health risk monitoring and real-time feedback to patients and care providers.

Mathematical Justification of Expression-Based Pathway Activation Scoring (PAS).

Although modeling of activation kinetics for various cell signaling pathways has reached a high grade of sophistication and thoroughness, most such kinetic models still remain of rather limited practical value for biomedicine. Nevertheless, recent advancements have been made in application of signaling pathway science for real needs of prescription of the most effective drugs for individual patients. The methods for such prescription evaluate the degree of pathological changes in the signaling machinery based on two types of data: first, on the results of high-throughput gene expression profiling, and second, on the molecular pathway graphs that reflect interactions between the pathway members. For example, our algorithm OncoFinder evaluates the activation of molecular pathways on the basis of gene/protein expression data in the objects of the interest.Yet, the question of assessment of the relative importance for each gene product in a molecular pathway remains unclear unless one call for the methods of parameter sensitivity /stiffness analysis in the interactomic kinetic models of signaling pathway activation in terms of total concentrations of each gene product.Here we show two principal points: 1. First, the importance coefficients for each gene in pathways that were obtained using the extremely time- and labor-consuming stiffness analysis of full-scaled kinetic models generally differ from much easier-to-calculate expression-based pathway activation score (PAS) not more than by 30%, so the concept of PAS is kinetically justified. 2. Second, the use of pathway-based approach instead of distinct gene analysis, due to the law of large numbers, allows restoring the correlation between the similar samples that were examined using different transcriptome investigation techniques.

Oncofinder, a new method for the analysis of intracellular signaling pathway activation using transcriptomic data.

We propose a new biomathematical method, OncoFinder, for both quantitative and qualitative analysis of the intracellular signaling pathway activation (SPA). This method is universal and may be used for the analysis of any physiological, stress, malignancy and other perturbed conditions at the molecular level. In contrast to the other existing techniques for aggregation and generalization of the gene expression data for individual samples, we suggest to distinguish the positive/activator and negative/repressor role of every gene product in each pathway. We show that the relative importance of each gene product in a pathway can be assessed using kinetic models for "low-level" protein interactions. Although the importance factors for the pathway members cannot be so far established for most of the signaling pathways due to the lack of the required experimental data, we showed that ignoring these factors can be sometimes acceptable and that the simplified formula for SPA evaluation may be applied for many cases. We hope that due to its universal applicability, the method OncoFinder will be widely used by the researcher community.