Search

In this work, we address the challenge of reconstructing the complete 12-lead ECG signal from incomplete parts of it. We focus on two main scenarii: (i) reconstructing missing signal segments within an ECG lead and (ii) recovering missing leads from a single-lead. We propose a model with a U-Net architecture trained on a novel objective function to address the reconstruction problem. This function incorporates both spatial and temporal aspects of the ECG by combining the distance in amplitude between the reconstructed and real signals with the signal trend. Through comprehensive assessments using both a real-life dataset and a publicly accessible one, we demonstrate that the proposed approach consistently outperforms state-of-the-art methods based on generative adversarial networks and a CopyPaste strategy. Our proposed model demonstrates superior performance in standard distortion metrics and preserves critical ECG characteristics, particularly the P, Q, R, S, and T wave coordinates. Two emerging clinical applications emphasize the relevance of our work. The first is the increasing need to digitize paper-stored ECGs for utilization in AI-based applications (automatic annotation and risk-quantification), often limited to digital ECG complete 10s recordings. The second is the widespread use of wearable devices that record ECGs but typically capture only a small subset of the 12 standard leads. In both cases, a non-negligible amount of information is lost or not recorded, which our approach aims to recover to overcome these limitations.

Deep learning (DL) techniques have shown unprecedented success when applied to images, waveforms, and text. Generally, when the sample size ($N$) is much bigger than the number of features ($d$), DL often outperforms other machine learning (ML) techniques, often through the use of Convolutional Neural Networks (CNNs). However, in many bioinformatics fields (including metagenomics), we encounter the opposite situation where $d$ is significantly greater than $N$. In these situations, applying DL techniques would lead to severe overfitting. Here we aim to improve classification of various diseases with metagenomic data through the use of CNNs. For this we proposed to represent metagenomic data as images. The proposed Met2Img approach relies on taxonomic and t-SNE embeddings to transform abundance data into "synthetic images". We applied our approach to twelve benchmark data sets including more than 1400 metagenomic samples. Our results show significant improvements over the state-of-the-art algorithms (Random Forest (RF), Support Vector Machine (SVM)). We observe that the integration of phylogenetic information alongside abundance data improves classification. The proposed approach is not only important in classification setting but also allows to visualize complex metagenomic data. The Met2Img is implemented in Python., Comment: The annual French Conference in Machine Learning (CAp 2018) - La Conf\'erence sur l'Apprentissage automatique (CAp), June 2018, Rouen, France. Oral presentation C12. http://cap2018.litislab.fr/Programme-en.html 10 pages, 7 figures, 4 tables

Deep learning (DL) techniques have had unprecedented success when applied to images, waveforms, and texts to cite a few. In general, when the sample size (N) is much greater than the number of features (d), DL outperforms previous machine learning (ML) techniques, often through the use of convolution neural networks (CNNs). However, in many bioinformatics ML tasks, we encounter the opposite situation where d is greater than N. In these situations, applying DL techniques (such as feed-forward networks) would lead to severe overfitting. Thus, sparse ML techniques (such as LASSO e.g.) usually yield the best results on these tasks. In this paper, we show how to apply CNNs on data which do not have originally an image structure (in particular on metagenomic data). Our first contribution is to show how to map metagenomic data in a meaningful way to 1D or 2D images. Based on this representation, we then apply a CNN, with the aim of predicting various diseases. The proposed approach is applied on six different datasets including in total over 1000 samples from various diseases. This approach could be a promising one for prediction tasks in the bioinformatics field., Comment: Accepted at NIPS 2017 Workshop on Machine Learning for Health (https://ml4health.github.io/2017/); In Proceedings of the NIPS ML4H 2017 Workshop in Long Beach, CA, USA

Well-designed medical decision support system (DSS) have been shown to improve health care quality. However, before they can be used in real clinical situations, these systems must be extensively tested, to ensure that they conform to the clinical guidelines (CG) on which they are based. Existing methods cannot be used for the systematic testing of all possible test cases. We describe here a new exhaustive dynamic verification method. In this method, the DSS is considered to be a black box, and the Quinlan C4.5 algorithm is used to build a decision tree from an exhaustive set of DSS input vectors and outputs. This method was successfully used for the testing of a medical DSS relating to chronic diseases: the ASTI critiquing module for type 2 diabetes.

Our view of the microbial world and its impact on human health is changing radically with the ability to sequence uncultured or unculturable microbes sampled directly from their habitats, ability made possible by fast and cheap next generation sequencing technologies. Such recent developments represents a paradigmatic shift in the analysis of habitat biodiversity, be it the human, soil or ocean microbiome. We review here some research examples and results that indicate the importance of the microbiome in our lives and then discus some of the challenges faced by metagenomic experiments and the subsequent analysis of the generated data. We then analyze the economic and social impact on genomic-medicine and research in both developing and developed countries. We support the idea that there are significant benefits in building capacities for developing high-level scientific research in metagenomics in developing countries. Indeed, the notion that developing countries should wait for developed countries to make advances in science and technology that they later import at great cost has recently been challenged.

We study the stress profile of an ordered two-dimensional packing of beads in response to the application of a vertical overload localized at its top surface. Disorder is introduced through the Coulombic friction between the grains which gives some indeterminacy and allows the choice of one constrained random number per grain in the calculation of the contact forces. The so-called `multi-agent' technique we use, lets us deal with systems as large as $1000\times1000$ grains. We show that the average response profile has a double peaked structure. At large depth $z$, the position of these peaks grows with $cz$, while their widths scales like $\sqrt{Dz}$. $c$ and $D$ are analogous to `propagation' and `diffusion' coefficients. Their values depend on that of the friction coefficient $\mu$. At small $\mu$, we get $c_0-c \propto \mu$ and $D \propto \mu^\beta$, with $\beta \sim 2.5$, which means that the peaks get closer and wider as the disorder gets larger. This behavior is qualitatively what was predicted in a model where a stochastic relation between the stress components is assumed., Comment: 7 pages, 7 figures, accepted version to Europhys. Lett