We present Tesla-Rapture, a gesture recognition system for sparse point clouds generated by mmWave Radars. State of the art gesture recognition models are either too resource consuming or not sufficiently accurate for the integration into real-life scenarios using wearable or constrained equipment such as IoT devices (e.g., Raspberry PI), XR hardware (e.g., HoloLens), or smart-phones. To tackle this issue, we have developed Tesla, a Message Passing Neural Network (MPNN) graph convolution approach for mmWave radar point clouds. The model outperforms the state of the art on three datasets in terms of accuracy while reducing the computational complexity and, hence, the execution time. In particular, the approach, is able to predict a gesture almost 8 times faster than the most accurate competitor. Our performance evaluation in different scenarios (environments, angles, distances) shows that Tesla generalizes well and improves the accuracy up to 20% in challenging scenarios, such as a through-wall setting and sensing at extreme angles. Utilizing Tesla, we develop Tesla-Rapture, a real-time implementation using a mmWave Radar on a Raspberry PI 4 and evaluate its accuracy and time-complexity. We also publish the source code, the trained models, and the implementation of the model for embedded devices.

BACKGROUND: Molecular interaction networks summarize complex biological processes as graphs, whose structure is informative of biological function at multiple scales. Simultaneously, omics technologies measure the variation or activity of genes, proteins, or metabolites across individuals or experimental conditions. Integrating the complementary viewpoints of biological networks and omics data is an important task in bioinformatics, but existing methods treat networks as discrete structures, which are intrinsically difficult to integrate with continuous node features or activity measures. Graph neural networks map graph nodes into a low-dimensional vector space representation, and can be trained to preserve both the local graph structure and the similarity between node features. RESULTS: We studied the representation of transcriptional, protein-protein and genetic interaction networks in E. coli and mouse using graph neural networks. We found that such representations explain a large proportion of variation in gene expression data, and that using gene expression data as node features improves the reconstruction of the graph from the embedding. We further proposed a new end-to-end Graph Feature Auto-Encoder framework for the prediction of node features utilizing the structure of the gene networks, which is trained on the feature prediction task, and showed that it performs better at predicting unobserved node features than regular MultiLayer Perceptrons. When applied to the problem of imputing missing data in single-cell RNAseq data, the Graph Feature Auto-Encoder utilizing our new graph convolution layer called FeatGraphConv outperformed a state-of-the-art imputation method that does not use protein interaction information, showing the benefit of integrating biological networks and omics data with our proposed approach. CONCLUSION: Our proposed Graph Feature Auto-Encoder framework is a powerful approach for integrating and exploiting the close relation between molecular interaction networks and functional genomics data.

MOTIVATION: Recently, it has become feasible to generate large-scale, multi-tissue gene expression data, where expression profiles are obtained from multiple tissues or organs sampled from dozens to hundreds of individuals. When traditional clustering methods are applied to this type of data, important information is lost, because they either require all tissues to be analyzed independently, ignoring dependencies and similarities between tissues, or to merge tissues in a single, monolithic dataset, ignoring individual characteristics of tissues. RESULTS: We developed a Bayesian model-based multi-tissue clustering algorithm, revamp, which can incorporate prior information on physiological tissue similarity, and which results in a set of clusters, each consisting of a core set of genes conserved across tissues as well as differential sets of genes specific to one or more subsets of tissues. Using data from seven vascular and metabolic tissues from over 100 individuals in the STockholm Atherosclerosis Gene Expression (STAGE) study, we demonstrate that multi-tissue clusters inferred by revamp are more enriched for tissue-dependent protein-protein interactions compared to alternative approaches. We further demonstrate that revamp results in easily interpretable multi-tissue gene expression associations to key coronary artery disease processes and clinical phenotypes in the STAGE individuals. AVAILABILITY AND IMPLEMENTATION: Revamp is implemented in the Lemon-Tree software, available at https://github.com/eb00/lemon-tree. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The lasso and elastic net linear regression models impose a double-exponential prior distribution on the model parameters to achieve regression shrinkage and variable selection, allowing the inference of robust models from large data sets. However, there has been limited success in deriving estimates for the full posterior distribution of regression coefficients in these models, due to a need to evaluate analytically intractable partition function integrals. Here, the Fourier transform is used to express these integrals as complex-valued oscillatory integrals over "regression frequencies". This results in an analytic expansion and stationary phase approximation for the partition functions of the Bayesian lasso and elastic net, where the non-differentiability of the double-exponential prior has so far eluded such an approach. Use of this approximation leads to highly accurate numerical estimates for the expectation values and marginal posterior distributions of the regression coefficients, and allows for Bayesian inference of much higher dimensional models than previously possible.

Mapping gene expression as a quantitative trait using whole genome-sequencing and transcriptome analysis allows to discover the functional consequences of genetic variation. We developed a novel method and ultra-fast software Findr for higly accurate causal inference between gene expression traits using cis-regulatory DNA variations as causal anchors, which improves current methods by taking into consideration hidden confounders and weak regulations. Findr outperformed existing methods on the DREAM5 Systems Genetics challenge and on the prediction of microRNA and transcription factor targets in human lymphoblastoid cells, while being nearly a million times faster. Findr is publicly available at https://github.com/lingfeiwang/findr.

Module network inference is an established statistical method to reconstruct co-expression modules and their upstream regulatory programs from integrated multi-omics datasets measuring the activity levels of various cellular components across different individuals, experimental conditions or time points of a dynamic process. We have developed Lemon-Tree, an open-source, platform-independent, modular, extensible software package implementing state-of-the-art ensemble methods for module network inference. We benchmarked Lemon-Tree using large-scale tumor datasets and showed that Lemon-Tree algorithms compare favorably with state-of-the-art module network inference software. We also analyzed a large dataset of somatic copy-number alterations and gene expression levels measured in glioblastoma samples from The Cancer Genome Atlas and found that Lemon-Tree correctly identifies known glioblastoma oncogenes and tumor suppressors as master regulators in the inferred module network. Novel candidate driver genes predicted by Lemon-Tree were validated using tumor pathway and survival analyses. Lemon-Tree is available from http://lemon-tree.googlecode.com under the GNU General Public License version 2.0.

BACKGROUND: The Kruskal-Wallis test is a popular non-parametric statistical test for identifying expression quantitative trait loci (eQTLs) from genome-wide data due to its robustness against variations in the underlying genetic model and expression trait distribution, but testing billions of marker-trait combinations one-by-one can become computationally prohibitive. RESULTS: We developed kruX, an algorithm implemented in Matlab, Python and R that uses matrix multiplications to simultaneously calculate the Kruskal-Wallis test statistic for several millions of marker-trait combinations at once. KruX is more than ten thousand times faster than computing associations one-by-one on a typical human dataset. We used kruX and a dataset of more than 500k SNPs and 20k expression traits measured in 102 human blood samples to compare eQTLs detected by the Kruskal-Wallis test to eQTLs detected by the parametric ANOVA and linear model methods. We found that the Kruskal-Wallis test is more robust against data outliers and heterogeneous genotype group sizes and detects a higher proportion of non-linear associations, but is more conservative for calling additive linear associations. CONCLUSION: kruX enables the use of robust non-parametric methods for massive eQTL mapping without the need for a high-performance computing infrastructure and is freely available from http://krux.googlecode.com.

MOTIVATION: Probabilistic motif detection requires a multi-step approach going from the actual de novo regulatory motif finding up to a tedious assessment of the predicted motifs. MotifSuite, a user-friendly web interface streamlines this analysis flow. Its core consists of two post-processing procedures that allow prioritizing the motif detection output. The tools offered by MotifSuite are built around the well-established motif detection tool MotifSampler and can also be used in combination with any other probabilistic motif detection tool. Elaborate guidelines on each of its applications have been provided. AVAILABILITY: http://homes.esat.kuleuven.be/bioi_marchal/MotifSuite/Index.htm

MOTIVATION: Transcriptional regulatory network inference methods have been studied for years. Most of them rely on complex mathematical and algorithmic concepts, making them hard to adapt, re-implement or integrate with other methods. To address this problem, we introduce a novel method based on a minimal statistical model for observing transcriptional regulatory interactions in noisy expression data, which is conceptually simple, easy to implement and integrate in any statistical software environment and equally well performing as existing methods. RESULTS: We developed a method to infer regulatory interactions based on a model where transcription factors (TFs) and their targets are both differentially expressed in a gene-specific, critical sample contrast, as measured by repeated two-way t-tests. Benchmarking on standard Escherichia coli and yeast reference datasets showed that this method performs equally well as the best existing methods. Analysis of the predicted interactions suggested that it works best to infer context-specific TF-target interactions which only co-express locally. We confirmed this hypothesis on a dataset of >1000 normal human tissue samples, where we found that our method predicts highly tissue-specific and functionally relevant interactions, whereas a global co-expression method only associates general TFs to non-specific biological processes. AVAILABILITY: A software tool called TwixTrix is available from http://twixtrix.googlecode.com. SUPPLEMENTARY INFORMATION: Supplementary Material is available from http://www.roslin.ed.ac.uk/tom-michoel/supplementary-data. CONTACT: [email protected].

Abstract Summary: Network motifs in integrated molecular networks represent functional relationships between distinct data types. They aggregate to form dense topological structures corresponding to functional modules which cannot be detected by traditional graph clustering algorithms. We developed CyClus3D, a Cytoscape plugin for clustering composite three-node network motifs using a 3D spectral clustering algorithm. Availability: Via the Cytoscape plugin manager or http://bioinformatics.psb.ugent.be/software/details/CyClus3D. Contact: [email protected] Supplementary Information: Supplementary data are available at Bioinformatics online.

Many algorithms have been proposed to learn transcription regulatory networks from gene expression data. Bayesian networks have obtained promising results, in particular, the module network method. The genes in a module share a regulation program (regression tree), consisting of a set of parents and conditional probability distributions. Hence, the method significantly decreases the search space of models and consequently avoids overfitting. The regulation program of a module is normally learned by a deterministic search algorithm, which performs a series of greedy operations to maximize the Bayesian score. The major shortcoming of the deterministic search algorithm is that its result may only represent one of several possible regulation programs. In order to account for the model uncertainty, we propose a regression tree-based Gibbs sampling algorithm for learning regulation programs in module networks. The novelty of this work is that a set of tree operations is defined for generating new regression trees from a given tree and we show that the set of tree operations is sufficient to generate a well mixing Gibbs sampler even in large data sets. The effectiveness of our algorithm is demonstrated by the experiments in synthetic data and real biological data.

MOTIVATION: Cancer is a complex disease, triggered by mutations in multiple genes and pathways. There is a growing interest in the application of systems biology approaches to analyze various types of cancer-related data to understand the overwhelming complexity of changes induced by the disease. RESULTS: We reconstructed a regulatory module network using gene expression, microRNA expression and a clinical parameter, all measured in lymphoblastoid cell lines derived from patients having aggressive or non-aggressive forms of prostate cancer. Our analysis identified several modules enriched in cell cycle-related genes as well as novel functional categories that might be linked to prostate cancer. Almost one-third of the regulators predicted to control the expression levels of the modules are microRNAs. Several of them have already been characterized as causal in various diseases, including cancer. We also predicted novel microRNAs that have never been associated to this type of tumor. Furthermore, the condition-dependent expression of several modules could be linked to the value of a clinical parameter characterizing the aggressiveness of the prostate cancer. Taken together, our results help to shed light on the consequences of aggressive and non-aggressive forms of prostate cancer. AVAILABILITY: The complete regulatory network is available as an interactive supplementary web site at the following URL: http://bioinformatics.psb.ugent.be/webtools/pronet/.

MOTIVATION: The solution of high-dimensional inference and prediction problems in computational biology is almost always a compromise between mathematical theory and practical constraints, such as limited computational resources. As time progresses, computational power increases but well-established inference methods often remain locked in their initial suboptimal solution. RESULTS: We revisit the approach of Segal et al. to infer regulatory modules and their condition-specific regulators from gene expression data. In contrast to their direct optimization-based solution, we use a more representative centroid-like solution extracted from an ensemble of possible statistical models to explain the data. The ensemble method automatically selects a subset of most informative genes and builds a quantitatively better model for them. Genes which cluster together in the majority of models produce functionally more coherent modules. Regulators which are consistently assigned to a module are more often supported by literature, but a single model always contains many regulator assignments not supported by the ensemble. Reliably detecting condition-specific or combinatorial regulation is particularly hard in a single optimum but can be achieved using ensemble averaging. AVAILABILITY: All software developed for this study is available from http://bioinformatics.psb.ugent.be/software.

MOTIVATION: Over the last decade, a large variety of clustering algorithms have been developed to detect coregulatory relationships among genes from microarray gene expression data. Model-based clustering approaches have emerged as statistically well-grounded methods, but the properties of these algorithms when applied to large-scale data sets are not always well understood. An in-depth analysis can reveal important insights about the performance of the algorithm, the expected quality of the output clusters, and the possibilities for extracting more relevant information out of a particular data set. RESULTS: We have extended an existing algorithm for model-based clustering of genes to simultaneously cluster genes and conditions, and used three large compendia of gene expression data for Saccharomyces cerevisiae to analyze its properties. The algorithm uses a Bayesian approach and a Gibbs sampling procedure to iteratively update the cluster assignment of each gene and condition. For large-scale data sets, the posterior distribution is strongly peaked on a limited number of equiprobable clusterings. A GO annotation analysis shows that these local maxima are all biologically equally significant, and that simultaneously clustering genes and conditions performs better than only clustering genes and assuming independent conditions. A collection of distinct equivalent clusterings can be summarized as a weighted graph on the set of genes, from which we extract fuzzy, overlapping clusters using a graph spectral method. The cores of these fuzzy clusters contain tight sets of strongly coexpressed genes, while the overlaps exhibit relations between genes showing only partial coexpression. AVAILABILITY: GaneSh, a Java package for coclustering, is available under the terms of the GNU General Public License from our website at http://bioinformatics.psb.ugent.be/software

BACKGROUND: In recent years, several authors have used probabilistic graphical models to learn expression modules and their regulatory programs from gene expression data. Despite the demonstrated success of such algorithms in uncovering biologically relevant regulatory relations, further developments in the area are hampered by a lack of tools to compare the performance of alternative module network learning strategies. Here, we demonstrate the use of the synthetic data generator SynTReN for the purpose of testing and comparing module network learning algorithms. We introduce a software package for learning module networks, called LeMoNe, which incorporates a novel strategy for learning regulatory programs. Novelties include the use of a bottom-up Bayesian hierarchical clustering to construct the regulatory programs, and the use of a conditional entropy measure to assign regulators to the regulation program nodes. Using SynTReN data, we test the performance of LeMoNe in a completely controlled situation and assess the effect of the methodological changes we made with respect to an existing software package, namely Genomica. Additionally, we assess the effect of various parameters, such as the size of the data set and the amount of noise, on the inference performance. RESULTS: Overall, application of Genomica and LeMoNe to simulated data sets gave comparable results. However, LeMoNe offers some advantages, one of them being that the learning process is considerably faster for larger data sets. Additionally, we show that the location of the regulators in the LeMoNe regulation programs and their conditional entropy may be used to prioritize regulators for functional validation, and that the combination of the bottom-up clustering strategy with the conditional entropy-based assignment of regulators improves the handling of missing or hidden regulators. CONCLUSION: We show that data simulators such as SynTReN are very well suited for the purpose of developing, testing and improving module network algorithms. We used SynTReN data to develop and test an alternative module network learning strategy, which is incorporated in the software package LeMoNe, and we provide evidence that this alternative strategy has several advantages with respect to existing methods.