Rxivist logo

Rxivist combines preprints from bioRxiv with data from Twitter to help you find the papers being discussed in your field. Currently indexing 62,725 bioRxiv papers from 278,303 authors.

Most downloaded bioRxiv papers, since beginning of last month

in category bioinformatics

6,115 results found. For more information, click each entry to expand.

1: GeneWalk identifies relevant gene functions for a biological context using network representation learning
more details view paper

Posted to bioRxiv 05 Sep 2019

GeneWalk identifies relevant gene functions for a biological context using network representation learning
1,843 downloads bioinformatics

Robert Ietswaart, Benjamin M Gyori, John A Bachman, Peter K Sorger, L. Stirling Churchman

The primary bottleneck in high-throughput genomics experiments is identifying the most important genes and their relevant functions from a list of gene hits. Existing methods such as Gene Ontology (GO) enrichment analysis provide insight at the gene set level. For individual genes, GO annotations are static and biological context can only be added by manual literature searches. Here, we introduce GeneWalk (github.com/churchmanlab/genewalk), a method that identifies individual genes and their relevant functions under a particular experimental condition. After automatic assembly of an experiment-specific gene regulatory network, GeneWalk quantifies the similarity between vector representations of each gene and its GO annotations through representation learning, yielding annotation significance scores that reflect their functional relevance for the experimental context. We demonstrate the use of GeneWalk analysis of RNA-seq and nascent transcriptome (NET-seq) data from human cells and mouse brains, validating the methodology. By performing gene- and condition-specific functional analysis that converts a list of genes into data-driven hypotheses, GeneWalk accelerates the interpretation of high-throughput genetics experiments.

2: Pan-cancer classifications of tumor histological images using deep learning
more details view paper

Posted to bioRxiv 26 Jul 2019

Pan-cancer classifications of tumor histological images using deep learning
1,639 downloads bioinformatics

Javad Noorbakhsh, Saman Farahmand, Mohammad Soltanieh-ha, Sandeep Namburi, Kourosh Zarringhalam, Jeff Chuang

Histopathological images are essential for the diagnosis of cancer type and selection of optimal treatment. However, the current clinical process of manual inspection of images is time consuming and prone to intra- and inter-observer variability. Here we show that key aspects of cancer image analysis can be performed by deep convolutional neural networks (CNNs) across a wide spectrum of cancer types. In particular, we implement CNN architectures based on Google Inception v3 transfer learning to analyze 27815 H&E slides from 23 cohorts in The Cancer Genome Atlas in studies of tumor/normal status, cancer subtype, and mutation status. For 19 solid cancer types we are able to classify tumor/normal status of whole slide images with extremely high AUCs (0.995±0.008). We are also able to classify cancer subtypes within 10 tissue types with AUC values well above random expectations (micro-average 0.87±0.1). We then perform a cross-classification analysis of tumor/normal status across tumor types. We find that classifiers trained on one type are often effective in distinguishing tumor from normal in other cancer types, with the relationships among classifiers matching known cancer tissue relationships. For the more challenging problem of mutational status, we are able to classify TP53 mutations in three cancer types with AUCs from 0.65-0.80 using a fully-trained CNN, and with similar cross-classification accuracy across tissues. These studies demonstrate the power of CNNs for not only classifying histopathological images in diverse cancer types, but also for revealing shared biology between tumors. We have made software available at: https://github.com/javadnoorb/HistCNN

3: Efficient de novo assembly of eleven human genomes using PromethION sequencing and a novel nanopore toolkit
more details view paper

Posted to bioRxiv 26 Jul 2019

Efficient de novo assembly of eleven human genomes using PromethION sequencing and a novel nanopore toolkit
1,041 downloads bioinformatics

Kishwar Shafin, Trevor Pesout, Ryan Lorig-Roach, Marina Haukness, Hugh E Olsen, Colleen Bosworth, Joel Armstrong, Kristof Tigyi, Nicholas Maurer, Sergey Koren, Fritz J. Sedlazeck, Tobias Marschall, Simon Mayes, Vania Costa, Justin M Zook, Kelvin J Liu, Duncan Kilburn, Melanie Sorensen, Katy M Munson, Mitchell R. Vollger, Evan E Eichler, Sofie Salama, David Haussler, Richard E. Green, Mark Akeson, Adam Phillippy, Karen H Miga, Paolo Carnevali, Miten Jain, Benedict Paten

Present workflows for producing human genome assemblies from long-read technologies have cost and production time bottlenecks that prohibit efficient scaling to large cohorts. We demonstrate an optimized PromethION nanopore sequencing method for eleven human genomes. The sequencing, performed on one machine in nine days, achieved an average 63x coverage, 42 Kb read N50, 90% median read identity and 6.5x coverage in 100 Kb+ reads using just three flow cells per sample. To assemble these data we introduce new computational tools: Shasta - a de novo long read assembler, and MarginPolish & HELEN - a suite of nanopore assembly polishing algorithms. On a single commercial compute node Shasta can produce a complete human genome assembly in under six hours, and MarginPolish & HELEN can polish the result in just over a day, achieving 99.9% identity (QV30) for haploid samples from nanopore reads alone. We evaluate assembly performance for diploid, haploid and trio-binned human samples in terms of accuracy, cost, and time and demonstrate improvements relative to current state-of-the-art methods in all areas. We further show that addition of proximity ligation (Hi-C) sequencing yields near chromosome-level scaffolds for all eleven genomes.

4: Modular and efficient pre-processing of single-cell RNA-seq
more details view paper

Posted to bioRxiv 17 Jun 2019

Modular and efficient pre-processing of single-cell RNA-seq
937 downloads bioinformatics

Páll Melsted, A. Sina Booeshaghi, Fan Gao, Eduardo da Veiga Beltrame, Lambda Lu, Kristján Eldjárn Hjorleifsson, Jase Gehring, Lior Pachter

Analysis of single-cell RNA-seq data begins with pre-processing of sequencing reads to generate count matrices. We investigate algorithm choices for the challenges of pre-processing, and describe a workflow that balances efficiency and accuracy. Our workflow is based on the kallisto (<https://pachterlab.github.io/kallisto/>) and bustools (<https://bustools.github.io/>) programs, and is near-optimal in speed and memory. The workflow is modular, and we demonstrate its flexibility by showing how it can be used for RNA velocity analyses. Documentation and tutorials for using the kallisto | bus workflow are available at <https://www.kallistobus.tools/>.

5: Improved metagenomic analysis with Kraken 2
more details view paper

Posted to bioRxiv 07 Sep 2019

Improved metagenomic analysis with Kraken 2
915 downloads bioinformatics

Derrick E Wood, Jennifer Lu, Benjamin Langmead

Although Kraken's k-mer-based approach provides fast taxonomic classification of metagenomic sequence data, its large memory requirements can be limiting for some applications. Kraken 2 improves upon Kraken 1 by reducing memory usage by 85%, allowing greater amounts of reference genomic data to be used, while maintaining high accuracy and increasing speed five-fold. Kraken 2 also introduces a translated search mode, providing increased sensitivity in viral metagenomics analysis.

6: Molecular Cross-Validation for Single-Cell RNA-seq
more details view paper

Posted to bioRxiv 30 Sep 2019

Molecular Cross-Validation for Single-Cell RNA-seq
910 downloads bioinformatics

Joshua Batson, Loic Royer, James T Webber

Single-cell RNA sequencing enables researchers to study the gene expression of individual cells. However, in high-throughput methods the portrait of each individual cell is noisy, representing thousands of the hundreds of thousands of mRNA molecules originally present. While many methods for denoising single-cell data have been proposed, a principled procedure for selecting and calibrating the best method for a given dataset has been lacking. We present "molecular cross-validation," a statistically principled and data-driven approach for estimating the accuracy of any denoising method without the need for ground-truth. We validate this approach for three denoising methods--principal component analysis, network diffusion, and a deep autoencoder--on a dataset of deeply-sequenced neurons. We show that molecular cross-validation correctly selects the optimal parameters for each method and identifies the best method for the dataset.

7: Assessment of computational methods for the analysis of single-cell ATAC-seq data
more details view paper

Posted to bioRxiv 18 Aug 2019

Assessment of computational methods for the analysis of single-cell ATAC-seq data
899 downloads bioinformatics

Huidong Chen, Caleb Lareau, Tommaso Andreani, Michael E. Vinyard, Sara P. Garcia, Kendell Clement, Miguel A. Andrade-Navarro, Jason Daniel Buenrostro, Luca Pinello

Background: Recent innovations in single-cell Assay for Transposase Accessible Chromatin using sequencing (scATAC-seq) enable profiling of the epigenetic landscape of thousands of individual cells. scATAC-seq data analysis presents unique methodological challenges. scATAC-seq experiments sample DNA, which, due to low copy numbers (diploid in humans) lead to inherent data sparsity (1-10% of peaks detected per cell) compared to transcriptomic (scRNA-seq) data (20-50% of expressed genes detected per cell). Such challenges in data generation emphasize the need for informative features to assess cell heterogeneity at the chromatin level. Results: We present a benchmarking framework that was applied to 10 computational methods for scATAC-seq on 13 synthetic and real datasets from different assays, profiling cell types from diverse tissues and organisms. Methods for processing and featurizing scATAC-seq data were evaluated by their ability to discriminate cell types when combined with common unsupervised clustering approaches. We rank evaluated methods and discuss computational challenges associated with scATAC-seq analysis including inherently sparse data, determination of features, peak calling, the effects of sequencing coverage and noise, and clustering performance. Running times and memory requirements are also discussed. Conclusions: This reference summary of scATAC-seq methods offers recommendations for best practices with consideration for both the non-expert user and the methods developer. Despite variation across methods and datasets, SnapATAC, Cusanovich2018, and cisTopic outperform other methods in separating cell populations of different coverages and noise levels in both synthetic and real datasets. Notably, SnapATAC was the only method able to analyze a large dataset (> 80,000 cells).

8: Deep Learning-Based Point-Scanning Super-Resolution Imaging
more details view paper

Posted to bioRxiv 21 Aug 2019

Deep Learning-Based Point-Scanning Super-Resolution Imaging
838 downloads bioinformatics

Linjing Fang, Fred Monroe, Sammy Weiser Novak, Lyndsey Kirk, Cara R. Schiavon, Seungyoon B. Yu, Tong Zhang, Melissa Wu, Kyle Kastner, Yoshiyuki Kubota, Zhao Zhang, Gulcin Pekkurnaz, John Mendenhall, Kristen Harris, Jeremy Howard, Uri Manor

Point scanning imaging systems (e.g. scanning electron or laser scanning confocal microscopes) are perhaps the most widely used tools for high resolution cellular and tissue imaging. Like all other imaging modalities, the resolution, speed, sample preservation, and signal-to-noise ratio (SNR) of point scanning systems are difficult to optimize simultaneously. In particular, point scanning systems are uniquely constrained by an inverse relationship between imaging speed and pixel resolution. Here we show these limitations can be mitigated via the use of deep learning-based super-sampling of undersampled images acquired on a point-scanning system, which we termed point-scanning super-resolution (PSSR) imaging. Oversampled, high SNR ground truth images acquired on scanning electron or Airyscan laser scanning confocal microscopes were 'crappified' to generate semi-synthetic training data for PSSR models that were then used to restore real-world undersampled images. Remarkably, our EM PSSR model could restore undersampled images acquired with different optics, detectors, samples, or sample preparation methods in other labs. PSSR enabled previously unattainable 2nm resolution images with our serial block face scanning electron microscope system. For fluorescence, we show that undersampled confocal images combined with a multiframe PSSR model trained on Airyscan timelapses facilitates Airyscan-equivalent spatial resolution and SNR with ~100x lower laser dose and 16x higher frame rates than corresponding high-resolution acquisitions. In conclusion, PSSR facilitates point-scanning image acquisition with otherwise unattainable resolution, speed, and sensitivity.

9: Telomere-to-telomere assembly of a complete human X chromosome
more details view paper

Posted to bioRxiv 16 Aug 2019

Telomere-to-telomere assembly of a complete human X chromosome
740 downloads bioinformatics

Karen H Miga, Sergey Koren, Arang Rhie, Mitchell R. Vollger, Ariel Gershman, Andrey Bzikadze, Shelise Brooks, Edmund Howe, David Porubsky, Glennis A. Logsdon, Valerie A Schneider, Tamara Potapova, Jonathan Wood, William Chow, Joel Armstrong, Jeanne Fredrickson, Evgenia Pak, Kristof Tigyi, Milinn Kremitzki, Christopher Markovic, Valerie Maduro, Amalia Dutra, Gerard G Bouffard, Alexander M Chang, Nancy F Hansen, Françoisen Thibaud-Nissen, Anthony D Schmitt, Jon-Matthew Belton, Siddarth Selvaraj, Megan Y. Dennis, Daniela C Soto, Ruta Sahasrabudhe, Gulhan Kaya, Josh Quick, Nicholas J Loman, Nadine Holmes, Matthew Loose, Urvashi Surti, Rosa ana Risques, Tina A. Graves Lindsay, Robert Fulton, Ira Hall, Benedict Paten, Kerstin Howe, Winston Timp, Alice Young, James C. Mullikin, Pavel A Pevzner, Jennifer E. Gerton, Beth A Sullivan, Evan E Eichler, Adam M Phillippy

After nearly two decades of improvements, the current human reference genome (GRCh38) is the most accurate and complete vertebrate genome ever produced. However, no one chromosome has been finished end to end, and hundreds of unresolved gaps persist [1][1],[2][2]. The remaining gaps include ribosomal rDNA arrays, large near-identical segmental duplications, and satellite DNA arrays. These regions harbor largely unexplored variation of unknown consequence, and their absence from the current reference genome can lead to experimental artifacts and hide true variants when re-sequencing additional human genomes. Here we present a de novo human genome assembly that surpasses the continuity of GRCh38 [2][2], along with the first gapless, telomere-to-telomere assembly of a human chromosome. This was enabled by high-coverage, ultra-long-read nanopore sequencing of the complete hydatidiform mole CHM13 genome, combined with complementary technologies for quality improvement and validation. Focusing our efforts on the human X chromosome [3][3], we reconstructed the ∼2.8 megabase centromeric satellite DNA array and closed all 29 remaining gaps in the current reference, including new sequence from the human pseudoautosomal regions and cancer-testis ampliconic gene families (CT-X and GAGE). This complete chromosome X, combined with the ultra-long nanopore data, also allowed us to map methylation patterns across complex tandem repeats and satellite arrays for the first time. These results demonstrate that finishing the human genome is now within reach and will enable ongoing efforts to complete the remaining human chromosomes. [1]: #ref-1 [2]: #ref-2 [3]: #ref-3

10: A Systematic Evaluation of Single Cell RNA-Seq Analysis Pipelines
more details view paper

Posted to bioRxiv 19 Mar 2019

A Systematic Evaluation of Single Cell RNA-Seq Analysis Pipelines
674 downloads bioinformatics

Beate Vieth, Swati Parekh, Christoph Ziegenhain, Wolfgang Enard, Ines Hellmann

The recent rapid spread of single cell RNA sequencing (scRNA-seq) methods has created a large variety of experimental and computational pipelines for which best practices have not been established, yet. Here, we use simulations based on five scRNA-seq library protocols in combination with nine realistic differential expression (DE) setups to systematically evaluate three mapping, four imputation, seven normalisation and four differential expression testing approaches resulting in ∼ 3,000 pipelines, allowing us to also assess interactions among pipeline steps. We find that choices of normalisation and library preparation protocols have the biggest impact on scRNA-seq analyses. Specifically, we find that library preparation determines the ability to detect symmetric expression differences, while normalisation dominates pipeline performance in asymmetric DE-setups. Finally, we illustrate the importance of informed choices by showing that a good scRNA-seq pipeline can have the same impact on detecting a biological signal as quadrupling the sample size.

11: Direct Comparative Analysis of 10X Genomics Chromium and Smart-seq2
more details view paper

Posted to bioRxiv 22 Apr 2019

Direct Comparative Analysis of 10X Genomics Chromium and Smart-seq2
663 downloads bioinformatics

Xiliang Wang, Yao He, Qiming Zhang, Xianwen Ren, Zemin Zhang

Single cell RNA sequencing (scRNA-seq) is widely used for profiling transcriptomes of individual cells. The droplet-based 10X Genomics Chromium (10X) approach and the plate-based Smart-seq2 full-length method are two frequently-used scRNA-seq platforms, yet there are only a few thorough and systematic comparisons of their advantages and limitations. Here, by directly comparing the scRNA-seq data by the two platforms from the same samples of CD45- cells, we systematically evaluated their features using a wide spectrum of analysis. Smart-seq2 detected more genes in a cell, especially low abundance transcripts as well as alternatively spliced transcripts, but captured higher proportion of mitochondrial genes. The composite of Smart-seq2 data also resembled bulk RNA-seq data better. For 10X-based data, we observed higher noise for mRNA in the low expression level. Despite the poly(A) enrichment, approximately 10-30% of all detected transcripts by both platforms were from non-coding genes, with lncRNA accounting for a higher proportion in 10X. 10X-based data displayed more severe dropout problem, especially for genes with lower expression levels. However, 10X-data can better detect rare cell types given its ability to cover a large number of cells. In addition, each platform detected different sets of differentially expressed genes between cell clusters, indicating the complementary nature of these technologies. Our comprehensive benchmark analysis offers the basis for selecting the optimal scRNA-seq strategy based on the objectives of each study.

12: PICRUSt2: An improved and extensible approach for metagenome inference
more details view paper

Posted to bioRxiv 15 Jun 2019

PICRUSt2: An improved and extensible approach for metagenome inference
653 downloads bioinformatics

Gavin M Douglas, Vincent J. Maffei, Jesse Zaneveld, Svetlana N Yurgel, James R Brown, Christopher M Taylor, Curtis Huttenhower, Morgan G I Langille

One major limitation of microbial community marker gene sequencing is that it does not provide direct information on the functional composition of sampled communities. Here, we present PICRUSt2, which expands the capabilities of the original PICRUSt method to predict approximate functional potential of a community based on marker gene sequencing profiles. This updated method and implementation includes several improvements over the previous algorithm: an expanded database of gene families and reference genomes, a new approach now compatible with any OTU-picking or denoising algorithm, novel phenotype predictions, and novel fungal reference databases that enable predictions from 18S rRNA gene and internal transcribed spacer amplicon data. Upon evaluation, PICRUSt2 was more accurate than PICRUSt1 and other current approaches and also more flexible to allow the addition of custom reference databases. Last, we demonstrate the utility of PICRUSt2 by identifying potential disease-associated microbial functional signatures based on 16S rRNA gene sequencing of ileal biopsies collected from a cohort of human subjects with inflammatory bowel disease. PICRUSt2 is freely available at: https://github.com/picrust/picrust2.

13: Visualizing Structure and Transitions for Biological Data Exploration
more details view paper

Posted to bioRxiv 24 Mar 2017

Visualizing Structure and Transitions for Biological Data Exploration
639 downloads bioinformatics

Kevin R Moon, David van Dijk, Zheng Wang, Scott Gigante, Daniel B Burkhardt, William S Chen, Kristina Yim, Antonia van den Elzen, Matthew J Hirn, Ronald R. Coifman, Natalia B Ivanova, Guy Wolf, Smita Krishnaswamy

With the advent of high-throughput technologies measuring high-dimensional biological data, there is a pressing need for visualization tools that reveal the structure and emergent patterns of data in an intuitive form. We present PHATE, a visualization method that captures both local and global nonlinear structure in data by an information-geometric distance between datapoints. We perform extensive comparison between PHATE and other tools on a variety of artificial and biological datasets, and find that it consistently preserves a range of patterns in data including continual progressions, branches, and clusters. We define a manifold preservation metric DEMaP to show that PHATE produces quantitatively better denoised embeddings than existing visualization methods. We show that PHATE is able to gain unique insight from a newly generated scRNA-seq dataset of human germ layer differentiation. Here, PHATE reveals a dynamic picture of the main developmental branches in unparalleled detail, including the identification of three novel subpopulations. Finally, we show that PHATE is applicable to a wide variety of datatypes including mass cytometry, single-cell RNA-sequencing, Hi-C, and gut microbiome data, where it can generate interpretable insights into the underlying systems.

14: Comparison of long-read sequencing technologies in the hybrid assembly of complex bacterial genomes
more details view paper

Posted to bioRxiv 26 Jan 2019

Comparison of long-read sequencing technologies in the hybrid assembly of complex bacterial genomes
618 downloads bioinformatics

Nicola De Maio, Liam P. Shaw, Alasdair Hubbard, Sophie George, Nick Sanderson, Jeremy Swann, Ryan Wick, Manal AbuOun, Emma Stubberfield, Sarah J Hoosdally, Derrick W Crook, Timothy E. A. Peto, Anna E Sheppard, Mark J. Bailey, Daniel S Read, Muna F. Anjum, A Sarah Walker, Nicole Stoesser, The REHAB consortium

Illumina sequencing allows rapid, cheap and accurate whole genome bacterial analyses, but short reads (<300 bp) do not usually enable complete genome assembly. Long read sequencing greatly assists with resolving complex bacterial genomes, particularly when combined with short-read Illumina data (hybrid assembly); however, it is not clear how different long-read sequencing methods impact on assembly accuracy. Relative automation of the assembly process is also crucial to facilitating high-throughput complete bacterial genome reconstruction, avoiding multiple bespoke filtering and data manipulation steps. In this study, we compared hybrid assemblies for 20 bacterial isolates, including two reference strains, using Illumina sequencing and long reads from either Oxford Nanopore Technologies (ONT) or from SMRT Pacific Biosciences (PacBio) sequencing platforms. We chose isolates from the Enterobacteriaceae family, as these frequently have highly plastic, repetitive genetic structures and complete genome reconstruction for these species is relevant for a precise understanding of the epidemiology of antimicrobial resistance. We de novo assembled genomes using the hybrid assembler Unicycler and compared different read processing strategies. Both strategies facilitate high-quality genome reconstruction. Combining ONT and Illumina reads fully resolved most genomes without additional manual steps, and at a lower cost per isolate in our setting. Automated hybrid assembly is a powerful tool for complete and accurate bacterial genome assembly.

15: End-to-end differentiable learning of protein structure
more details view paper

Posted to bioRxiv 14 Feb 2018

End-to-end differentiable learning of protein structure
577 downloads bioinformatics

Mohammed AlQuraishi

Accurate prediction of protein structure is one of the central challenges of biochemistry. Despite significant progress made by co-evolution methods to predict protein structure from signatures of residue-residue coupling found in the evolutionary record, a direct and explicit mapping between protein sequence and structure remains elusive, with no substantial recent progress. Meanwhile, rapid developments in deep learning, which have found remarkable success in computer vision, natural language processing, and quantum chemistry raise the question of whether a deep learning based approach to protein structure could yield similar advancements. A key ingredient of the success of deep learning is the reformulation of complex, human-designed, multi-stage pipelines with differentiable models that can be jointly optimized end-to-end. We report the development of such a model, which reformulates the entire structure prediction pipeline using differentiable primitives. Achieving this required combining four technical ideas: (1) the adoption of a recurrent neural architecture to encode the internal representation of protein sequence, (2) the parameterization of (local) protein structure by torsional angles, which provides a way to reason over protein conformations without violating the covalent chemistry of protein chains, (3) the coupling of local protein structure to its global representation via recurrent geometric units, and (4) the use of a differentiable loss function to capture deviations between predicted and experimental structures. To our knowledge this is the first end-to-end differentiable model for learning of protein structure. We test the effectiveness of this approach using two challenging tasks: the prediction of novel protein folds without the use of co-evolutionary information, and the prediction of known protein folds without the use of structural templates. On the first task the model achieves state-of-the-art performance, even when compared to methods that rely on co-evolutionary data. On the second task the model is competitive with methods that use experimental protein structures as templates, achieving 3-7Å accuracy despite being template-free. Beyond protein structure prediction, end-to-end differentiable models of proteins represent a new paradigm for learning and modeling protein structure, with potential applications in docking, molecular dynamics, and protein design.

16: Deep learning for brains?: Different linear and nonlinear scaling in UK Biobank brain images vs. machine-learning datasets
more details view paper

Posted to bioRxiv 06 Sep 2019

Deep learning for brains?: Different linear and nonlinear scaling in UK Biobank brain images vs. machine-learning datasets
572 downloads bioinformatics

Marc-Andre Schulz, Thomas Yeo, Joshua Vogelstein, Janaina Mourao-Miranada, Jakob Kather, Konrad Kording, B.T. Thomas Yeo, Danilo Bzdok

In recent years, deep learning has unlocked unprecedented success in various domains, especially in image, text, and speech processing. These breakthroughs may hold promise for neuroscience and especially for brain-imaging investigators who start to analyze thousands of participants. However, deep learning is only beneficial if the data have nonlinear relationships and if they are exploitable at currently available sample sizes. We systematically profiled the performance of deep models, kernel models, and linear models as a function of sample size on UK Biobank brain images against established machine learning references. On MNIST and Zalando Fashion, prediction accuracy consistently improved when escalating from linear models to shallow-nonlinear models, and further improved when switching to deep-nonlinear models. The more observations were available for model training, the greater the performance gain we saw. In contrast, using structural or functional brain scans, simple linear models performed on par with more complex, highly parameterized models in age/sex prediction across increasing sample sizes. In fact, linear models kept improving as the sample size approached ~10,000 participants. Our results indicate that the increase in performance of linear models with additional data does not saturate at the limit of current feasibility. Yet, nonlinearities of common brain scans remain largely inaccessible to both kernel and deep learning methods at any examined scale.

17: Robustness and applicability of functional genomics tools on scRNA-seq data
more details view paper

Posted to bioRxiv 01 Sep 2019

Robustness and applicability of functional genomics tools on scRNA-seq data
543 downloads bioinformatics

Christian H Holland, Jovan Tanevski, Jan Gleixner, Manu P. Kumar, Elisabetta Mereu, Javier Perales-Paton, Brian A. Joughin, Oliver Stegle, Douglas A Lauffenburger, Holger Heyn, Bence Szalai, Julio Saez-Rodriguez

Many tools have been developed to extract functional and mechanistic insight from bulk transcriptome profiling data. With the advent of single-cell RNA sequencing (scRNA-seq), it is in principle possible to do such an analysis for single cells. However, scRNA-seq data has characteristics such as drop-out events, low library sizes and a comparatively large number of samples/cells. It is thus not clear if functional genomics tools established for bulk sequencing can be applied to scRNA-seq in a meaningful way. To address this question, we performed benchmark studies on in silico and in vitro single-cell RNA-seq data. We included the bulk-RNA tools PROGENy, GO enrichment and DoRothEA that estimate pathway and transcription factor (TF) activities, respectively, and compared them against the tools AUCell and metaVIPER, designed for scRNA-seq. For the in silico study we simulated single cells from TF/pathway perturbation bulk RNA-seq experiments. Our simulation strategy guarantees that the information of the original perturbation is preserved while resembling the characteristics of scRNA-seq data. We complemented the in silico data with in vitro scRNA-seq data upon CRISPR-mediated knock-out. Our benchmarks on both the simulated and real data revealed comparable performance to the original bulk data. Additionally, we showed that the TF and pathway activities preserve cell-type specific variability by analysing a mixture sample sequenced with 13 scRNA-seq different protocols. Our analyses suggest that bulk functional genomics tools can be applied to scRNA-seq data, outperforming dedicated single cell tools. Furthermore we provide a benchmark for further methods development by the community.

18: scClassify: hierarchical classification of cells
more details view paper

Posted to bioRxiv 23 Sep 2019

scClassify: hierarchical classification of cells
524 downloads bioinformatics

Yingxin Lin, Yue Cao, Hani J Kim, Agus Salim, Terence P Speed, Dave Lin, Pengyi Yang

Cell type identification is a key computational challenge in single-cell RNA-sequencing (scRNA-seq) data. To capitalize on the large collections of well-annotated scRNA-seq datasets, we present scClassify, a hierarchical classification framework based on ensemble learning. scClassify can identify cells from published scRNA-seq datasets more accurately and more finely than in the original publications. We also estimate the cell number needed for accurate classification anywhere in a cell type hierarchy.

19: The art of using t-SNE for single-cell transcriptomics
more details view paper

Posted to bioRxiv 25 Oct 2018

The art of using t-SNE for single-cell transcriptomics
520 downloads bioinformatics

Dmitry Kobak, Philipp Berens

Single-cell transcriptomics yields ever growing data sets containing RNA expression levels for thousands of genes from up to millions of cells. Common data analysis pipelines include a dimensionality reduction step for visualising the data in two dimensions, most frequently performed using t-distributed stochastic neighbour embedding (t-SNE). It excels at revealing local structure in high-dimensional data, but naive applications often suffer from severe shortcomings, e.g. the global structure of the data is not represented accurately. Here we describe how to circumvent such pitfalls, and develop a protocol for creating more faithful t-SNE visualisations. It includes PCA initialisation, a high learning rate, and multi-scale similarity kernels; for very large data sets, we additionally use exaggeration and downsampling-based initialisation. We use published single-cell RNA-seq data sets to demonstrate that this protocol yields superior results compared to the naive application of t-SNE.

20: Performance of neural network basecalling tools for Oxford Nanopore sequencing
more details view paper

Posted to bioRxiv 07 Feb 2019

Performance of neural network basecalling tools for Oxford Nanopore sequencing
498 downloads bioinformatics

Ryan R Wick, Louise M Judd, Kathryn Holt

Basecalling, the computational process of translating raw electrical signal to nucleotide sequence, is of critical importance to the sequencing platforms produced by Oxford Nanopore Technologies (ONT). Here we examine the performance of different basecalling tools, looking at accuracy at the level of bases within individual reads and at majority-rules consensus basecalls in an assembly. We also investigate some additional aspects of basecalling: training using a taxon-specific dataset, using a larger neural network model and improving consensus basecalls in an assembly by additional signal-level analysis with Nanopolish. Training basecallers on taxon-specific data results in a significant boost in consensus accuracy, mostly due to the reduction of errors in methylation motifs. A larger neural network is able to improve both read and consensus accuracy, but at a cost to speed. Improving consensus sequences ('polishing') with Nanopolish somewhat negates the accuracy differences in basecallers, but prepolish accuracy does have an effect on post-polish accuracy. Basecalling accuracy has seen significant improvements over the last two years. The current version of ONT's Guppy basecaller performs well overall, with good accuracy and fast performance. If higher accuracy is required, users should consider producing a custom model using a larger neural network and/or training data from the same species.

Previous page 1 2 3 4 5 . . . 306 Next page

Sign up for the Rxivist weekly newsletter! (Click here for more details.)