Artificial neural networks are popular for computer vision as they often give state-of-the-art performance, but are difficult to interpret because of their complexity. This black box modeling is especially troubling when the application concerns human well-being such as in medical image analysis or autonomous driving. In this work, we propose a technique called routing path visualization for capsule networks, which reveals how much of each region in an image is routed to each capsule. In turn, this technique can be used to interpret the entity that a given capsule detects, and speculate how the network makes a prediction. We demonstrate our new visualization technique on several real world datasets. Experimental results suggest that routing path visualization can precisely localize the predicted class from an image, even though the capsule networks are trained using just images and their respective class labels, without additional information defining the location of the class in the image.

BACKGROUND: Transcriptional regulation is a fundamental mechanism underlying biological functions. In recent years, a broad array of RNA-Seq tools have been used to measure transcription levels in biological experiments, in whole organisms, tissues, and at the single cell level. Collectively, this is a vast comparative dataset on transcriptional processes across organisms. Yet, due to technical differences between the studies (sequencing, experimental design, and analysis) extracting usable comparative information and conducting meta-analyses remains challenging. RESULTS: We introduce Comparative RNA-Seq Metadata Analysis Pipeline (CoRMAP), a meta-analysis tool to retrieve comparative gene expression data from any RNA-Seq dataset using de novo assembly, standardized gene expression tools and the implementation of OrthoMCL, a gene orthology search algorithm. It employs the use of orthogroup assignments to ensure the accurate comparison of gene expression levels between experiments and species. Here we demonstrate the use of CoRMAP on two mouse brain transcriptomes with similar scope, that were collected several years from each other using different sequencing technologies and analysis methods. We also compare the performance of CoRMAP with a functional mapping tool, previously published. CONCLUSION: CoRMAP provides a framework for the meta-analysis of RNA-Seq data from divergent taxonomic groups. This method facilitates the retrieval and comparison of gene expression levels from published data sets using standardized assembly and analysis. CoRMAP does not rely on reference genomes and consequently facilitates direct comparison between diverse studies on a range of organisms.

UNLABELLED: Dragon Promoter Mapper (DPM) is a tool to model promoter structure of co-regulated genes using methodology of Bayesian networks. DPM exploits an exhaustive set of motif features (such as motif, its strand, the order of motif occurrence and mutual distance between the adjacent motifs) and generates models from the target promoter sequences, which may be used to (1) detect regions in a genomic sequence which are similar to the target promoters or (2) to classify other promoters as similar or not to the target promoter group. DPM can also be used for modelling of enhancers and silencers. AVAILABILITY: http://defiant.i2r.a-star.edu.sg/projects/BayesPromoter/ CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Manual for using DPM web server is provided at http://defiant.i2r.a-star.edu.sg/projects/BayesPromoter/html/manual/manual.htm.

MOTIVATION: Histone proteins play important roles in chromosomal functions. They are significantly evolutionarily conserved across species, which suggests similarity in their transcription regulation. The abundance of experimental data on histone promoters provides an excellent background for the evaluation of computational methods. Our study addresses the issue of how well computational analysis can contribute to unveiling the biologically relevant content of promoter regions for a large number of mammalian histone genes taken across several species, and suggests the consensus promoter models of different histone groups. RESULTS: This is the first study to unveil the detailed promoter structures of all five mammalian histone groups and their subgroups. This is also the most comprehensive computational analysis of histone promoters performed to date. The most exciting fact is that the results correlate very well with the biologically known facts and experimental data. Our analysis convincingly demonstrates that computational approach can significantly contribute to elucidation of promoter content (identification of biologically relevant signals) complementing tedious wet-lab experiments. We believe that this type of analysis can be easily applied to other functional gene classes, thus providing a general framework for modelling promoter groups. These results also provide the basis to hunt for genes co-regulated with histone genes across mammalian genomes.