PaperPlayer biorxiv bioinformatics

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.11.04.368803v1?rss=1

Authors: Zhou, H., Shrikumar, A., Kundaje, A.

Abstract:

Predictive models that map double-stranded regulatory DNA to molecular signals of regulatory activity should, in principle, produce identical predictions regardless of whether the sequence of the forward strand or its reverse complement (RC) is supplied as input. Unfortunately, standard convolutional neural network architectures can produce highly divergent predictions across strands, even when the training set is augmented with RC sequences. Two strategies have emerged in the literature to enforce this symmetry: conjoined a.k.a. "siamese" architectures where the model is run in parallel on both strands & predictions are combined, and RC parameter sharing or RCPS where weight sharing ensures that the response of the model is equivariant across strands. However, systematic benchmarks are lacking, and neither architecture has been adapted to base-resolution signal profile prediction tasks. In this work, we extend conjoined and RCPS models to signal profile prediction, and introduce a strong baseline: a standard model (trained on RC augmented data) that is converted to a conjoined model only after it has been trained, which we call a "post-hoc" conjoined model. We then conduct benchmarks on both binary and signal profile prediction. We find post-hoc conjoined models consistently perform as well as or better than models that were conjoined during training, and present a mathematical intuition for why. We also find that - despite its theoretical appeal - RCPS performs surprisingly poorly on certain tasks, in particular, signal profile prediction. In fact, RCPS can sometimes do worse than even standard models trained with RC data augmentation. We prove that the RCPS models can represent the solution learned by the conjoined models, implying that the poor performance of RCPS may be due to optimization difficulties. We therefore suggest that users interested in RC symmetry should default to post-hoc conjoined models as a reliable baseline before exploring RCPS. Code: https://github.com/hannahgz/BenchmarkRCStrategies

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.11.03.366237v1?rss=1

Authors: LAVENIER, D., Jodin, R., Cimadomo, R.

Abstract:

In this paper, we introduce a new combination of software and hardware PIM (Process-in-Memory) architecture to accelerate the variant calling genomic process. PIM translates into bringing data intensive calculations directly where the data is: within the DRAM, enhanced with thousands of processing units. The energy consumption, in large part due to data movement, is significantly lowered at a marginal additional hardware cost. Such design allows an unprecedented level of parallelism to process billions of short reads. Experiments on real PIM devices developed by the UPMEM company show significant speed-up compared to pure software implementation. The PIM solution also compared nicely to FPGA or GPU based acceleration bringing similar to twice the processing speed but most importantly being 5 to 8 times cheaper to deploy with up to 6 times less power consumption.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.11.03.366146v1?rss=1

Authors: Cao, K., Hong, Y., Wan, L.

Abstract:

Single-cell multi-omics sequencing data can provide a comprehensive molecular view of cells. However, effective approaches for the integrative analysis of such data are challenging. Although achieved state-of-the-art performance on single-cell multi-omics data integration and did not require any correspondence information, either among cells or among features, current manifold alignment based integrative methods are often limited by requiring that single-cell datasets be derived from the same underlying cellular structure. To overcome this limitation, we present Pamona, an algorithm that integrates heterogeneous single-cell multi-omics datasets with the aim of delineating and representing the shared and dataset-specific cellular structures. We formulate this task as a partial manifold alignment problem and develop a partial Gromov-Wasserstein optimal transport framework to solve it. Pamona identifies both shared and dataset-specific cells based on the computed probabilistic couplings of cells across datasets, and it aligns cellular modalities in a common low-dimensional space, while simultaneously preserving both shared and dataset-specific structures. Our framework can easily incorporate prior information, such as cell type annotations or cell-cell correspondence, to further improve alignment quality. Simulation studies and applications to four real data sets demonstrate that Pamona can accurately identify shared and dataset-specific cells, as well as faithfully recover and align cellular structures of heterogeneous single-cell modalities in the common space.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.11.02.365528v1?rss=1

Authors: Stephens, S. H., King, C. H., Watford, S., Patel, J., Dean, D. A., Koc, S., Xiao, N., Donaldson, E., Thompson, E., Purkayastha, A., Mazumder, R., Johanson, E., Keeney, J.

Abstract:

Background: The field of bioinformatics has grown at such a rapid pace that a gap in standardization exists when reporting an analysis. In response, the BioCompute project was created to standardize the type and method of information communicated when describing a bioinformatic analysis. Once the project became established, its goals shifted to broadening awareness and usage of BioCompute, and soliciting feedback from a larger audience. To address these goals, the BioCompute project collaborated with precisionFDA on a crowdsourced challenge that ran from May 2019 to October 2019. This challenge had a beginner track where participants submitted BCOs based on a pipeline of their choosing, and an advanced track where participants submitted applications supporting the creation of a BCO and verification of BCO conformance to specifications. Results: In total, there were 28 submissions to the beginner track (including submissions from a bioinformatics master's class at George Washington University) and three submissions to the advanced track. Three top performers were selected from the beginner track, while a single top performer was selected for the advanced track. In the beginner track, top performers differentiated themselves by submitting BCOs that included more than the minimally compliant content. Advanced track submissions were very impressive. They included a complete web application, a command line tool that produced a static result, and a dockerized container that automatically created the BCO as the tool was run. The ability to harmonize the correct function, a simple user experience, and the aesthetics of the tool interface differentiated the tools. Conclusions: Despite being new to the concept, most beginner track scores were high, indicating that most users understood the fundamental concepts of the BCO specification. Novice bioinformatics students were an ideal cohort for this Challenge because of their lack of familiarity with BioCompute, broad diversity of research interests, and motivation to submit high-quality work. This challenge was successful in introducing the BCO to a wider audience, obtaining feedback from that audience, and resulting in a tool novices may use for BCO creation and conformance. In addition, the BCO specification itself was improved based on feedback illustrating the utility of a "wisdom of the crowd" approach to standards development.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.11.02.360933v1?rss=1

Authors: Queiros, P., Delogu, F., Hickl, O., May, P., Wilmes, P.

Abstract:

Background: The past decades have seen a rapid development of the (meta-)omics fields, producing an unprecedented amount of data. Through the use of well-characterized datasets we can infer the role of previously functionally unannotated proteins from single organisms and consortia. In this context, protein function annotation allows the identification of regions of interest (i.e. domains) in protein sequences and the assignment of biological functions. Despite the existence of numerous tools, some challenges remain, specifically in terms of speed, flexibility, and reproducibility. In the era of big data it also becomes increasingly important to cease limiting our findings to a single reference, coalescing knowledge from different data sources, thus overcoming some limitations in overly relying on computationally generated data. Results: We implemented a protein annotation tool - Mantis, which uses text mining to integrate knowledge from multiple reference data sources into a single consensus-driven output. Mantis is flexible, allowing for total customization of the reference data used, adaptable, and reproducible across different research goals and user environments. We implemented a depth-first search algorithm for domain-specific annotation, which led to an average 0.038 increase in precision when compared to sequence-wide annotation. Mantis is fast, annotating an average genome in 25-40 minutes, whilst also outputting high-quality annotations (average coverage 81.4%, average precision 0.892). Conclusions: Mantis is a protein function annotation tool that produces high-quality consensus-driven protein annotations. It is easy to set up, customize, and use, scaling from single genomes to large metagenomes. Mantis is available under the MIT license available at https://github.com/PedroMTQ/mantis

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.11.02.360958v1?rss=1

Authors: Sirois, I., Caron, E., Kovalchik, K. A., Wessling, L., Saab, F., Ma, Q., Despault, J., Kubiniok, P., Hamelin, D., Faridi, P., Li, C., Purcell, A. W., Tognetti, M., Reiter, L., Bruderer, R., Lanoix, J., Bonneil, E., Courcelles, M., Thibault, P.

Abstract:

Immunopeptidomics refers to the science of investigating the composition and dynamics of peptides presented by major histocompatibility complex (MHC) class I and class II molecules using mass spectrometry (MS). Here, we aim to provide a technical report to any non-expert in the field wishing to establish and/or optimize an immunopeptidomic workflow with relatively limited computational knowledge and resources. To this end, we thoroughly describe step-by-step instructions to isolate MHC class I and II-associated peptides from various biological sources, including mouse and human biospecimens. Most notably, we created MhcVizPipe (MVP) (https://github.com/CaronLab/MhcVizPipe), a new and easy-to-use open-source software tool to rapidly assess the quality and the specific enrichment of immunopeptidomic datasets upon the establishment of new workflows. In fact, MVP enables intuitive visualization of multiple immunopeptidomic datasets upon testing sample preparation protocols and new antibodies for the isolation of MHC class I and II peptides. In addition, MVP enables the identification of unexpected binding motifs and facilitates the analysis of non-canonical MHC peptides. We anticipate that the experimental and bioinformatic resources provided herein will represent a great starting point for any non-expert and will therefore foster the accessibility and expansion of the field to ultimately boost its maturity and impact.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.11.02.364588v1?rss=1

Authors: Pavlova, A., Belova, V., Afasizhev, R., Bulusheva, I., Rebrikov, D., Korostin, D.

Abstract:

During the sequencing process, problems can occur with any device including the MGISEQ-2000 (DNBSEQ-G400) platform. We encountered a power outage that resulted in a temporary shutdown of a sequencer in the middle of the run. Since barcode reading in MGISEQ-2000 takes place at the end of the run, it was impossible to use non-demultiplexed raw data. We decided to completely use up the same cartridge with reagents and flow cell loaded with DNB and started a new run in a shortened cus-tom mode. We figured out how the MGISEQ-2000 converts preliminary data in .cal format into .fastq files and wrote a script named Runcer-Necromacer for merging .fastq files based on the analysis of their headers (available online: https://github.com/genomecenter/runcer-necromancer ). Read merging proved to be possible because the MGISEQ-2000 flow cell has a patterned structure and each DNB has invariable coordinates on it, regardless of its position on the flow cell stage. We demonstrated the correctness of data merging by comparing sample analysis results with previously obtained .fastq files for them. Thus, we confirmed that it is possible to restart the device and save both parts of the interrupted run.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.11.02.365841v1?rss=1

Authors: Nato, A. Q., Ata Ul Mustafa, H., Patel, R. K., Sexton, H. G., Moore, S. D., Denvir, J., Primerano, D. A., Risher, M.-L.

Abstract:

Introduction: Alcohol use disorder (AUD) is a complex, chronic psychiatric disease. AUD manifests as having uncontrollable drinking patterns with detrimental effects. Excessive alcohol intake in the form of binge drinking, which is common among adolescents and young adults, is associated with increased risk of developing AUD. Here, we analyze RNA-seq data from hippocampi of Sprague Dawley rats to investigate temporal changes in gene expression. We used a rodent model of binge drinking, i.e., adolescent intermittent ethanol (AIE), to identify candidate genes that may play a role in the chronic changes in brain function and contribute to the development of AUD. Methods: At postnatal day (PND) 30 (adolescence), rats received chronic intermittent ethanol (5g/kg intragastrically (i.g.) 10 times across 16 days). We extracted total RNA from rat hippocampal tissue that was collected at three time points. RNA was sequenced on an Illumina HiSeq 2000 platform. We processed RNA-seq data (TrimGalore), compiled gene counts (HTSeq), and performed differential expression analysis (DESeq2). The full rank list of genes and the differentially expressed genes (DEGs) with nominal p<0.05 were used as inputs for pathway-based enrichment analyses GSEA and Cytoscape, respectively, at 3 timepoints in the presence or absence of ethanol (i.e., 3 comparisons). For each of these comparisons, GSEA was used to identify genes involved in enriched Gene Ontology (GO) terms while Cytoscape and its apps were used to identify networks of genes and, subsequently, subnetworks (i.e., clusters) of genes enriched by higher interaction. From clusters containing 15 or more nodes, we prioritized genes related with particular functions, cell types, or diseases, which were present in GO Processes, Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathways, or Reactome Pathways. Results: Based on GSEA, the most impacted differential gene expression occurs within the potassium channels and receptor function but is heavily dependent on the time point at which the analysis occurs. Across both GSEA and Cytoscape pathway analyses, the most striking changes occur in genes that regulate neuroinflammatory processes and neuronal/synaptic remodeling and coincide with the enrichment of pathways involved in addiction processes. Conclusion: Identification of genes dysregulated by AIE may be useful in determining the underlying mechanism for acute and chronic effects of AIE exposure that contribute to neuronal remodeling and increased risk of developing AUD.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.11.02.365049v1?rss=1

Authors: Verma, S., Dwivedy, A., Kumar, N., Biswal, B. K.

Abstract:

Background: Over the past two decades, there has been a continued research on the role of small non-coding RNAs including microRNAs (miRNAs) in various diseases. Studies have shown that viruses modulate the host cellular machinery and hijack its metabolic and immune signaling pathways by miRNA mediated gene silencing. Given the immensity of coronavirus disease 19 (COVID-19) pandemic and the strong association of viral encoded miRNAs with their pathogenesis, it is important to study Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) miRNAs. Results: To address this unexplored area, we identified 8 putative novel miRNAs from SARS-CoV-2 genome and explored their possible human gene targets. A significant proportion of these targets populated key immune and metabolic pathways such as MAPK signaling pathway, maturity-onset diabetes of the young, Insulin signaling pathway, endocytosis, RNA transport, TGF-{beta} signaling pathway, to name a few. The data from this work is backed up by recently reported high-throughput transcriptomics datasets obtains from SARS-CoV-2 infected samples. Analysis of these datasets reveal that a significant proportion of the target human genes were down-regulated upon SARS-CoV-2 infection. Conclusions: The current study brings to light probable host metabolic and immune pathways susceptible to viral miRNA mediated silencing in a SARS-CoV-2 infection, and discusses its effects on the host pathophysiology.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.10.30.362590v1?rss=1

Authors: Gonzalez Buitron, M., Tunque Cahui, R. R., Garcia Rios, E., Hirsh, L., Fornasari, M. S., Parisi, G., Palopoli, N.

Abstract:

Conformational changes in RNA native ensembles are central to fulfill many of their biological roles. Systematic knowledge of the extent and possible modulators of this conformational diversity is desirable to better understand the relationship between RNA dynamics and function. We have developed CoDNaS-RNA as the first database of conformational diversity in RNA molecules. Known RNA structures are retrieved and clustered to identify alternative conformers of each molecule. Pairwise structural comparisons within each cluster allows to measure the variability of the molecule. Additional data on structural features, molecular interactions and functional annotations are provided. CoDNaS-RNA is implemented as a public resource that can be of much interest for computational and bench scientists alike.

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