Polina Mamoshina

Applications of Deep Learning in Biomedicine

Increases in throughput and installed base of biomedical research equipment led to a massive accumulation of -omics data known to be highly variable, high-dimensional, and sourced from multiple often incompatible data platforms. While this data may be useful for biomarker identification and drug discovery, the bulk of it remains underutilized. Deep neural networks (DNNs) are efficient algorithms based on the use of compositional layers of neurons, with advantages well matched to the challenges -omics data presents. While achieving state-of-the-art results and even surpassing human accuracy in many challenging tasks, the adoption of deep learning in biomedicine has been comparatively slow. Here, we discuss key features of deep learning that may give this approach an edge over other machine learning methods. We then consider limitations and review a number of applications of deep learning in biomedical studies demonstrating proof of concept and practical utility.

Deep Learning Applications for Predicting Pharmacological Properties of Drugs and Drug Repurposing Using Transcriptomic Data

Deep learning is rapidly advancing many areas of science and technology with multiple success stories in image, text, voice and video recognition, robotics, and autonomous driving. In this paper we demonstrate how deep neural networks (DNN) trained on large transcriptional response data sets can classify various drugs to therapeutic categories solely based on their transcriptional profiles. We used the perturbation samples of 678 drugs across A549, MCF-7, and PC-3 cell lines from the LINCS Project and linked those to 12 therapeutic use categories derived from MeSH. To train the DNN, we utilized both gene level transcriptomic data and transcriptomic data processed using a pathway activation scoring algorithm, for a pooled data set of samples perturbed with different concentrations of the drug for 6 and 24 hours. In both pathway and gene level classification, DNN achieved high classification accuracy and convincingly outperformed the support vector machine (SVM) model on every multiclass classification problem, however, models based on pathway level data performed significantly better. For the first time we demonstrate a deep learning neural net trained on transcriptomic data to recognize pharmacological properties of multiple drugs across different biological systems and conditions. We also propose using deep neural net confusion matrices for drug repositioning. This work is a proof of principle for applying deep learning to drug discovery and development.

Deep biomarkers of human aging: Application of deep neural networks to biomarker development

One of the major impediments in human aging research is the absence of a comprehensive and actionable set of biomarkers that may be targeted and measured to track the effectiveness of therapeutic interventions. In this study, we designed a modular ensemble of 21 deep neural networks (DNNs) of varying depth, structure and optimization to predict human chronological age using a basic blood test. To train the DNNs, we used over 60,000 samples from common blood biochemistry and cell count tests from routine health exams performed by a single laboratory and linked to chronological age and sex. The best performing DNN in the ensemble demonstrated 81.5 % epsilon-accuracy r = 0.90 with R2 = 0.80 and MAE = 6.07 years in predicting chronological age within a 10 year frame, while the entire ensemble achieved 83.5% epsilon-accuracy r = 0.91 with R2 = 0.82 and MAE = 5.55 years. The ensemble also identified the 5 most important markers for predicting human chronological age: albumin, glucose, alkaline phosphatase, urea and erythrocytes. To allow for public testing and evaluate real-life performance of the predictor, we developed an online system available at http://www.aging.ai. The ensemble approach may facilitate integration of multi-modal data linked to chronological age and sex that may lead to simple, minimally invasive, and affordable methods of tracking integrated biomarkers of aging in humans and performing cross-species feature importance analysis.

The cornucopia of meaningful leads: Applying deep adversarial autoencoders for new molecule development in oncology

Recent advances in deep learning and specifically in generative adversarial networks have demonstrated surprising results in generating new images and videos upon request even using natural language as input. In this paper we present the first application of generative adversarial autoencoders (AAE) for generating novel molecular fingerprints with a defined set of parameters. We developed a 7-layer AAE architecture with the latent middle layer serving as a discriminator. As an input and output the AAE uses a vector of binary fingerprints and concentration of the molecule. In the latent layer we also introduced a neuron responsible for growth inhibition percentage, which when negative indicates the reduction in the number of tumor cells after the treatment. To train the AAE we used the NCI-60 cell line assay data for 6252 compounds profiled on MCF-7 cell line. The output of the AAE was used to screen 72 million compounds in PubChem and select candidate molecules with potential anti-cancer properties. This approach is a proof of concept of an artificially-intelligent drug discovery engine, where AAEs are used to generate new molecular fingerprints with the desired molecular properties.

Design of efficient computational workflows for in silico drug repurposing

Here, we provide a comprehensive overview of the current status of in silico repurposing methods by establishing links between current technological trends, data availability and characteristics of the algorithms used in these methods. Using the case of the computational repurposing of fasudil as an alternative autophagy enhancer, we suggest a generic modular organization of a repurposing workflow. We also review 3D structure-based, similarity-based, inference-based and machine learning (ML)-based methods. We summarize the advantages and disadvantages of these methods to emphasize three current technical challenges. We finish by discussing current directions of research, including possibilities offered by new methods, such as deep learning.

In search for geroprotectors: in silico screening and in vitro validation of signalome-level mimetics of young healthy state

Populations in developed nations throughout the world are rapidly aging, and the search for geroprotectors, or anti-aging interventions, has never been more important. Yet while hundreds of geroprotectors have extended lifespan in animal models, none have yet been approved for widespread use in humans. GeroScope is a computational tool that can aid prediction of novel geroprotectors from existing human gene expression data. GeroScope maps expression differences between samples from young and old subjects to aging-related signaling pathways, then profiles pathway activation strength (PAS) for each condition. Known substances are then screened and ranked for those most likely to target differential pathways and mimic the young signalome. Here we used GeroScope and shortlisted ten substances, all of which have lifespan-extending effects in animal models, and tested 6 of them for geroprotective effects in senescent human fibroblast cultures. PD-98059, a highly selective MEK1 inhibitor, showed both life-prolonging and rejuvenating effects. Natural compounds like N-acetyl-L-cysteine, Myricetin and Epigallocatechin gallate also improved several senescence-associated properties and were further investigated with pathway analysis. This work not only highlights several potential geroprotectors for further study, but also serves as a proof-of-concept for GeroScope, Oncofinder and other PAS-based methods in streamlining drug prediction, repurposing and personalized medicine.

A review of the biomedical innovations for healthy longevity

Keywords: longevity ; aging ; biomarkers ; geroprotectors ; epigenetics ; transcriptomics Reference EPFL-ARTICLE-227576doi:10.18632/aging.101163View record in Web of Science Record created on 2017-05-01, modified on 2017-05-26

Use of deep neural network ensembles to identify embryonic-fetal transition markers: repression of COX7A1 in embryonic and cancer cells

Here we present the application of deep neural network (DNN) ensembles trained on transcriptomic data to identify the novel markers associated with the mammalian embryonic-fetal transition (EFT). Molecular markers of this process could provide important insights into regulatory mechanisms of normal development, epimorphic tissue regeneration and cancer. Subsequent analysis of the most significant genes behind the DNNs classifier on an independent dataset of adult-derived and human embryonic stem cell (hESC)-derived progenitor cell lines led to the identification of COX7A1 gene as a potential EFT marker. COX7A1, encoding a cytochrome C oxidase subunit, was up-regulated in post-EFT murine and human cells including adult stem cells, but was not expressed in pre-EFT pluripotent embryonic stem cells or their in vitro-derived progeny. COX7A1 expression level was observed to be undetectable or low in multiple sarcoma and carcinoma cell lines as compared to normal controls. The knockout of the gene in mice led to a marked glycolytic shift reminiscent of the Warburg effect that occurs in cancer cells. The DNN approach facilitated the elucidation of a potentially new biomarker of cancer and pre-EFT cells, the embryo-onco phenotype, which may potentially be used as a target for controlling the embryonic-fetal transition.