PaperPlayer biorxiv bioinformatics

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.04.283937v1?rss=1

Authors: Chen, C., Wu, T., Guo, Z., Cheng, J.

Abstract:

Deep learning has emerged as a revolutionary technology for protein residue-residue contact prediction since the 2012 CASP10 competition. Considerable advancements in the predictive power of the deep learning-based contact predictions have been achieved since then. However, little effort has been put into interpreting the black-box deep learning methods. Algorithms that can interpret the relationship between predicted contact maps and the internal mechanism of the deep learning architectures are needed to explore the essential components of contact inference and improve their explainability. In this study, we present an attention-based convolutional neural network for protein contact prediction, which consists of two attention mechanism-based modules: sequence attention and regional attention. Our benchmark results on the CASP13 free-modeling (FM) targets demonstrate that the two attention modules added on top of existing typical deep learning models exhibit a complementary effect that contributes to predictive improvements. More importantly, the inclusion of the attention mechanism provides interpretable patterns that contain useful insights into the key fold-determining residues in proteins. We expect the attention-based model can provide a reliable and practically interpretable technique that helps break the current bottlenecks in explaining deep neural networks for contact prediction.

Copy rights belong to original authors. Visit the link for more info

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.04.282269v1?rss=1

Authors: Cieslak, M., Cook, P. A., He, X., Yeh, F.-C., Dhollander, T., Adebimpe, A., Aguirre, G. K., Bassett, D. S., Betzel, R. F., Bourque, J., Cabral, L., Davatzikos, C., Detre, J., Earl, E., Elliott, M. A., Fadnavis, S., Fair, D. A., Foran, W., Fotiadis, P., Garyfallidis, E., Giesbrecht, B., Gur, R. C., Gur, R. E., Kelz, M., Keshavan, A., Larsen, B. S., Luna, B., Mackey, A. P., Milham, M., Oathes, D. J., Perrone, A., Pines, A. R., Roalf, D. R., Richie-Halford, A., Rokem, A., Sydnor, V., Tapera, T. M., Tooley, U. A., Vettel, J. M., Yeatman, J., Grafton, S. T., Satterthwaite, T. D.

Abstract:

Diffusion-weighted magnetic resonance imaging (dMRI) has become the primary method for non-invasively studying the organization of white matter in the human brain. While many dMRI acquisition sequences have been developed, they all sample q-space in order to characterize water diffusion. Numerous software platforms have been developed for processing dMRI data, but most work on only a subset of sampling schemes or implement only parts of the processing workflow. Reproducible research and comparisons across dMRI methods are hindered by incompatible software, diverse file formats, and inconsistent naming conventions. Here we introduce QSIPrep, an integrative software platform for the processing of diffusion images that is compatible with nearly all dMRI sampling schemes. Drawing upon a diverse set of software suites to capitalize upon their complementary strengths, QSIPrep automatically applies best practices for dMRI preprocessing, including denoising, distortion correction, head motion correction, coregistration, and spatial normalization. Throughout, QSIPrep provides both visual and quantitative measures of data quality as well as "glass-box" methods reporting. Taken together, these features facilitate easy implementation of best practices for processing of diffusion images while simultaneously ensuring reproducibility.

Copy rights belong to original authors. Visit the link for more info

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.04.281451v1?rss=1

Authors: Ferraro, L.

Abstract:

This paper is based on Computational Biology with an eye towards medicine. In fact our aim is to identify new indications for existing drugs with a computational approach using Systems Biology, in order to speed up the market entry, avoiding some phases of drug development (for example toxicity tests) and with a lower cost. The method that we will discuss exploits the concept of Random Walk in a heterogeneous network, formed by a drug-drug and disease-disease similarity and known drug-disease associations. This is done by combining different types of data in order to connect more effectively two drugs or two diseases with a similarity score. At the end we integrate the known associations between drugs and diseases with the purpose of finding similarity values of new couples.

Copy rights belong to original authors. Visit the link for more info

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.01.277079v1?rss=1

Authors: Yang, A., Yao, Y., Fang, X., Li, J., Xia, Y., Kwok, C. S. M., Lo, M. C. K., Siu, D. M. D., Tsia, K. K., Ho, J. W. K.

Abstract:

Advances in high throughput single-cell and spatial omic technologies have enabled the profiling of molecular expression and phenotypic properties of hundreds of thousands of individual cells in the context of their two dimensional (2D) or three dimensional (3D) spatial endogenous arrangement. However, current visualisation techniques do not allow for effective display and exploration of the single cell data in their spatial context. With the widespread availability of low-cost virtual reality (VR) gadgets, such as Google Cardboard, we propose that an immersive visualisation strategy is useful. We present starmapVR, a light-weight, cross-platform, web-based tool for visualising single-cell and spatial omic data. starmapVR supports a number of interaction methods, such as keyboard, mouse, wireless controller and voice control. The tool visualises single cells in a 3D space and each cell can be represented by a star plot (for molecular expression, phenotypic properties) or image (for single cell imaging). For spatial transcriptomic data, the 2D single cell expression data can be visualised alongside the histological image in a 2.5D format. The application of starmapVR is demonstrated through a series of case studies. Its scalability has been carefully evaluated across different platforms. starmapVR is freely accessible at https://holab-hku.github.io/starmapVR, with the corresponding source code available at https://github.com/holab-hku/starmapVR under the open source MIT license.

Copy rights belong to original authors. Visit the link for more info

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.04.282731v1?rss=1

Authors: Munz, M., Khodaygani, M., Aherrahrou, Z., Busch, H., Wohlers, I.

Abstract:

Mice are the most widely used animal model to study genotype to phenotype relationships. Inbred mice are genetically identical, which eliminates genetic heterogeneity and makes them particularly useful for genetic studies. Many different strains have been bred over decades and a vast amount of phenotypic data has been generated. In addition, lately, also whole genome sequencing-based genome-wide genotype data for many widely used inbred strains has been released. Here, we present an approach for in silico fine mapping that uses genotypic data of 37 inbred mouse strains together with phenotypic data provided by the user to propose candidate variants and genes for the phenotype under study. Public genome-wide genotype data covering more than 74 million variant sites is queried efficiently in real-time to provide those variants that are compatible with the observed phenotype differences between strains. Variants can be filtered by molecular consequences and by corresponding molecular impact. Candidate gene lists can be generated from variant lists on the fly. Fine mapping together with annotation or filtering of results is provided in a Bioconductor package called MouseFM. For albinism, MouseFM reports only one variant allele of moderate or high molecular impact that only albino mice share: a missense variant in the Tyr gene, reported previously to be causal for this phenotype. Performing in silico fine mapping for interfrontal bone formation in mice using four strains with and five strains without interfrontal bone results in 12 genes. Of these, three are related to skull shaping abnormality. Finally performing fine mapping for dystrophic cardiac calcification by comparing 8 strains showing the phenotype with 8 strains lacking it, we identify only one moderate impact variant in the known causal gene Abcc6. In summary, this illustrates the benefit of using MouseFM for candidate variant and gene identification.

Copy rights belong to original authors. Visit the link for more info

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.03.280958v1?rss=1

Authors: Schulz, T., Wittler, R., Rahmann, S., Hach, F., Stoye, J.

Abstract:

Motivation: Increasing amounts of individual genomes sequenced per species motivate the usage of pangenomic approaches. Pangenomes may be represented as graphical structures, e.g. compacted colored de Bruijn graphs, which offer a low memory usage and facilitate reference-free sequence comparisons. While sequence-to-graph mapping to graphical pangenomes has been studied for some time, no local alignment search tool in the vein of BLAST has been proposed yet. Results: We present a new heuristic method to find maximum scoring local alignments of a DNA query sequence to a pangenome represented as a compacted colored de Bruijn graph. Our approach additionally allows a comparison of similarity among sequences within the pangenome. We show that local alignment scores follow an exponential-tail distribution similar to BLAST scores, and we discuss how to estimate its parameters to separate local alignments representing sequence homology from spurious findings. An implementation of our method is presented, and its performance and usability are shown. Our approach scales sublinearly in running time and memory usage with respect to the number of genomes under consideration. This is an advantage over classical methods that do not make use of sequence similarity within the pangenome.

Copy rights belong to original authors. Visit the link for more info

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.02.279570v1?rss=1

Authors: Fellows Yates, J. A., Andrades Valtuena, A., Vagene, A. J., Cribdon, B., Velsko, I. M., Borry, M., Bravo-Lopez, M. J., Fernandez-Guerra, A., Green, E. J., Ramachandran, S. L., Heintzman, P. D., Spyrou, M. A., Hübner, A., Gancz, A. S., Hider, J., Allshouse, A. F., Warinner, C. G.

Abstract:

Ancient DNA and RNA are valuable data sources for a wide range of disciplines. Within the field of ancient metagenomics, the number of published genetic datasets has risen dramatically in recent years, and tracking this data for reuse is particularly important for large-scale ecological and evolutionary studies of individual microbial taxa, microbial communities, and metagenomic assemblages. AncientMetagenomeDir (archived at https://doi.org/10.5281/zenodo.3980834) is a collection of indices of published genetic data deriving from ancient microbial samples that provides basic, standardised metadata and accession numbers to allow rapid data retrieval from online repositories. These collections are community-curated and span multiple sub-disciplines in order to ensure adequate breadth and consensus in metadata definitions, as well as longevity of the database. Internal guidelines and automated checks to facilitate compatibility with established sequence-read archives and term-ontologies ensure consistency and interoperability for future meta-analyses. This collection will also assist in standardising metadata reporting for future ancient metagenomic studies.

Copy rights belong to original authors. Visit the link for more info

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.02.271098v1?rss=1

Authors: Sulit, A. K. L., Kolisnik, T., Frizelle, F. A., Purcell, R., Schmeier, S.

Abstract:

Background: The identification of functional processes taking place in microbiome communities augment traditional microbiome taxonomic studies, giving a more complete picture of interactions taking place within the community. While there are applications that perform functional annotation on metagenome or metatranscriptomes, very few of these are able to link taxonomic identity to function and are limited by their input types or databases used. Results: Here we present MetaFunc, a workflow which takes input reads, and from these 1) identifies species present in the microbiome sample and 2) provides gene ontology (GO) annotations associated with the species identified. MetaFunc can also provide a differential abundance analysis step comparing species between sample conditions. In addition, MetaFunc allows mapping of reads to a host genome, and separates these reads, before proceeding with the microbiome analyses. From the host reads, MetaFunc is able to identify host genes, perform differential gene expression analysis, and gene-set enrichment analysis. A final correlation analysis between microbial species and host genes can also be performed. Finally, MetaFunc builds an R shiny application that allows users to view and interact with the microbiome results. In this paper we show how MetaFunc can be applied to metatranscriptomic datasets of colorectal cancer. Conclusion: MetaFunc is a one-stop shop microbiome analysis pipeline that can identify taxonomies and their respective functional contributions in a microbiome sample through GO annotations. It can also analyse host reads in a microbiome sample, providing information on host gene expression, and allowing for correlations between the microbiome and host genes. MetaFunc comes with a user-friendly R shiny application that allows for easier visualisation and exploration of its results. MetaFunc is freely available through https://gitlab.com/schmeierlab/workflows/metafunc.git.

Copy rights belong to original authors. Visit the link for more info

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.03.269555v1?rss=1

Authors: Dexter, A., Thomas, S. A., Steven, R. T., Robinson, K. N., Taylor, A. J., Elia, E., Nikula, C., Campbell, A. D., Panina, Y., Najumudeen, A. K., Murta, T., Yan, B., Grabowski, P., Hamm, G., Swales, J., Gilmore, I., Yuneva, M., Goodwin, R. J. A., Barry, S., Sansom, O. J., Takats, Z., Bunch, J.

Abstract:

High dimensionality omics and hyperspectral imaging datasets present difficult challenges for feature extraction and data mining due to huge numbers of features that cannot be simultaneously examined. The sample numbers and variables of these methods are constantly growing as new technologies are developed, and computational analysis needs to evolve to keep up with growing demand. Current state of the art algorithms can handle some routine datasets but struggle when datasets grow above a certain size. We present a training deep learning via neural networks on non-linear dimensionality reduction, in particular t-distributed stochastic neighbour embedding (t-SNE), to overcome prior limitations of these methods.

Copy rights belong to original authors. Visit the link for more info

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.02.278804v1?rss=1

Authors: zeng, y., zhou, x., rao, j., lu, y., yang, y.

Abstract:

Recent advances in single-cell RNA sequencing (scRNA-seq) technologies provide a great opportunity to study gene expression at cellular resolution, and the scRNA-seq data has been routinely conducted to unfold cell heterogeneity and diversity. A critical step for the scRNA-seq analyses is to cluster the same type of cells, and many methods have been developed for cell clustering. However, existing clustering methods are limited to extract the representations from expression data of individual cells, while ignoring the high-order structural relations between cells. Here, we proposed a new method (GraphSCC) to cluster cells based on scRNA-seq data by accounting structural relations between cells through a graph convolutional network. The representation learned from the graph convolutional network, together with another representation output from a denoising autoencoder network, are optimized by a dual self-supervised module for better cell clustering. Extensive experiments indicate that GraphSCC model outperforms state-of-the-art methods in various evaluation metrics on both simulated and real datasets. Further visualizations show that GraphSCC provides representations for better intra-cluster compactness and inter-cluster separability.

Copy rights belong to original authors. Visit the link for more info