PaperPlayer biorxiv bioinformatics

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.18.303842v1?rss=1

Authors: Mammoliti, A., Smirnov, P., Nakano, M., Safikhani, Z., Ho, C., Beri, G., Haibe-Kains, B.

Abstract:

Reproducibility is essential to Open Science, as there is limited relevance for finding that cannot be reproduced by independent research groups, regardless of its validity. It is therefore crucial for scientists to describe their experiments in sufficient detail so they can be reproduced, challenged, and built upon. However, due to recent advances in the biological and computational sciences, it has become difficult to process, analyze, and share data with the community in a manner that is transparent. This has made reproducing research findings more challenging, with some researchers going as far as suggesting that the biomedical sciences are experiencing a "reproducibility crisis". To overcome these issues, we created a cloud-based platform called ORCESTRA (www.orcestra.ca), which provides a flexible framework for the reproducible processing of multimodal biomedical data. The platform enables processing of genomic and pharmacological profiles of cancer samples through the use of automated processing pipelines that are user-customizable, which are executed through Pachyderm, a data versioning and orchestration tool. ORCESTRA creates an integrated and fully documented data object known as a PharmacoSet (PSet), with a persistent identifier (DOI), that can be used and shared for future analyses using the Bioconductor PharmacoGx package.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.19.304923v1?rss=1

Authors: MENG, L., WANG, C., Shi, Y., LUO, Q.

Abstract:

There is a strong demand for the methods that can efficiently reconstruct biologically valid super-resolution intact genome 3D structures from sparse and noise single-cell Hi-C data. Here, we developed Single-Cell Chromosome Conformation Calculator (Si-C) within the Bayesian theory framework and applied this approach to reconstruct intact genome 3D structures from the single-cell Hi-C data of eight G1-phase haploid mouse ES cells. The inferred 100-kb and 10-kb structures consistently reproduce the known conserved features of chromatin organization revealed by independent imaging experiments. The analysis of the 10-kb resolution 3D structures revealed cell-to-cell varying domain structures in individual cells and hyperfine structures in domains, such as loops. An average of 0.2 contact reads per divided bin is sufficient for Si-C to obtain reliable structures. The valid super-resolution structures constructed by Si-C demonstrates the potential for visualizing and investigating interactions between all chromatin loci at genome scale in individual cells.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.18.278549v1?rss=1

Authors: Hershberger, J., Morales, N., Simoes, C. C., Ellerbrock, B., Bauchet, G., Mueller, L. A., Gore, M. A.

Abstract:

Visible and near-infrared (vis-NIRS) spectroscopy is a promising tool for increasing phenotyping throughput in plant breeding programs, but existing analysis software packages are not optimized for a breeding context. Additionally, commercial software options are often outside of budget constraints for some breeding and research programs. To that end, we developed an open-source R package, waves, for the streamlined analysis of spectral data with several cross-validation schemes to assess prediction accuracy. Waves is compatible with a wide range of spectrometer models and performs visualization, filtering, aggregation, cross-validation set formation, model training, and prediction functions for the association of vis-NIRS spectra with reference measurements. Furthermore, we have integrated this package into the Breedbase family of open-source databases, expanding the analysis capabilities of this growing digital ecosystem to a number of crop species. Taken together, the standalone and Breedbase versions of waves enhance the accessibility of tools for the analysis of spectral data during the plant breeding process.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.18.304329v1?rss=1

Authors: Pourreza Shahri, M., Lyon, K., Schearer, J., Kahanda, I.

Abstract:

The biomedical literature provides an extensive source of information in the form of unstructured text. One of the most important types of information hidden in biomedical literature is the relations between human proteins and their phenotypes, which, due to the exponential growth of publications, can remain hidden. This provides a range of opportunities for the development of computational methods to extract the biomedical relations from the unstructured text. In our previous work, we developed a supervised machine learning approach, called PPPred, for classifying the validity of a given sentence-level human protein-phenotype co-mention. In this work, we propose DeepPPPred, an ensemble classifier composed of PPPred and three deep neural network models: RNN, CNN, and BERT. Using an expanded gold-standard co-mention dataset, we demonstrate that the proposed ensemble method significantly outperforms its constituent components and provides a new state-of-the-art performance on classifying the co-mentions of human proteins and phenotype terms.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.17.301788v1?rss=1

Authors: Hahn, G., Lutz, S. M., Laha, N., Lange, C.

Abstract:

Penalized linear regression approaches that include an L1 term have become an important tool in day-to-day statistical data analysis. One prominent example is the least absolute shrinkage and selection operator (Lasso), though the class of L1 penalized regression operators also includes the fused and graphical Lasso, the elastic net, etc. Although the L1 penalty makes their objective function convex, it is not differentiable everywhere, motivating the development of proximal gradient algorithms such as Fista, the current gold standard in the literature. In this work, we take a different approach based on smoothing. The methodological contribution of our article is threefold: (1) We introduce a unified framework to compute closed-form smooth surrogates of a whole class of L1 penalized regression problems using Nesterov smoothing. The surrogates preserve the convexity of the original (unsmoothed) objective functions, are uniformly close to them, and have closed-form derivatives everywhere for efficient minimization via gradient descent; (2) We prove that the estimates obtained with the smooth surrogates can be made arbitrarily close to the ones of the original (unsmoothed) objective functions, and provide explicitly computable bounds on the accuracy of our estimates; (3) We propose an iterative algorithm to progressively smooth the L1 penalty which increases accuracy and is virtually free of tuning parameters. The proposed methodology is applicable to a large class of L1 penalized regression operators, including all the operators mentioned above. Using simulation studies, we compare our framework to current gold standards such as Fista, glmnet, gLasso, etc. Our simulation results suggest that our proposed smoothing framework provides estimates of equal or higher accuracy than the gold standards while keeping the aforementioned theoretical guarantees and having roughly the same asymptotic runtime scaling.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.17.301960v1?rss=1

Authors: Giacopuzzi, E., Popitsch, N., Taylor, J. C.

Abstract:

Background: Non-coding variants have emerged as important contributors to the pathogenesis of human diseases, not only as common susceptibility alleles but also as rare high-impact variants. Despite recent advances in the study of regulatory elements and the availability of specialized data collections, the systematic annotation of non-coding variants from genome sequencing remains challenging. Results: We integrated 24 data sources to develop a standardized collection of 2.4 million regulatory elements in the human genome, transcription factor binding sites, DNase peaks, ultra-conserved non-coding elements, and super-enhancers. Information on controlled gene(s), tissue(s) and associated phenotype(s) are provided for regulatory elements when possible. We also calculated a variation constraint metric for regulatory regions and showed that genes controlled by constrained regions are more likely to be disease-associated genes and essential genes from mouse knock-out screenings. Finally, we evaluated 16 non-coding impact prediction scores providing suggestions for variant prioritization. The companion tool allows for annotation of VCF files with information about the regulatory regions as well as non-coding prediction scores to inform variant prioritization. The proposed annotation framework was able to capture previously published disease-associated non-coding variants and its integration in a routine prioritization pipeline increased the number of candidate genes, including genes potentially correlated with patient phenotype, and established clinically relevant genes. Conclusion: We have developed a resource for the annotation and prioritization of regulatory variants in WGS analysis to support the discovery of candidate disease-associated variants in the non-coding genome.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.17.301879v1?rss=1

Authors: Ji, Y., Zhou, Z., Liu, H., Davuluri, R.

Abstract:

Deciphering the language of non-coding DNA is one of the fundamental problems in genome research. Gene regulatory code is highly complex due to the existence of polysemy and distant semantic relationship, which previous informatics methods often fail to capture especially in data-scarce scenarios. To address this challenge, we developed a novel pre-trained bidirectional encoder representation, named DNABERT, that forms global and transferrable understanding of genomic DNA sequences based on up and downstream nucleotide contexts. We show that the single pre-trained transformers model can simultaneously achieve state-of-the-art performance on many sequence predictions tasks, after easy fine-tuning using small task-specific data. Further, DNABERT enables direct visualization of nucleotide-level importance and semantic relationship within input sequences for better interpretability and accurate identification of conserved sequence motifs and functional genetic variants. Finally, we demonstrate that pre-trained DNABERT with human genome can even be readily applied to other organisms with exceptional performance.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.17.301911v1?rss=1

Authors: Lee, H., Kim, C., Jeong, J., Jung, K., Han, B.

Abstract:

Integrating scRNA-seq data of multiple donors is challenging. Multiple samples may exhibit strong heterogeneity and batch effects, which need to be properly corrected. Many previous methods focused on integrating multi-sample data in the cluster level, but it was challenging to quantitatively measure the benefit of integration. We present scIntegral, a scalable method to integrate hundreds of donors scRNA data. Our method aims to identify cell-types of the cells in a semi-supervised fashion using marker list information as prior. scIntegral is extremely efficient and takes only an hour to integrate ten thousand donor data, while fully accounting for heterogeneity with covariates. We quantify the benefit of multi-sample integration in terms of accuracy with respect to the gold standard cell labels, and prove that integrating multiple donors can significantly reduce the error rate in cell-type identification. scIntegral is more accurate than existing methods and can precisely identify very rare (<0.5%) cell populations, suggesting utilities for in-silico cell extraction. scIntegral is freely available at https://github.com/hanbin973/scIntegral.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.17.301515v1?rss=1

Authors: Sharma, A., Lysenko, A., Boroevich, K., Vans, E., Tsunoda, T.

Abstract:

Identifying smaller element or gene subsets from biological or other data types is an essential step in discovering underlying mechanisms. Statistical machine learning methods have played a key role in revealing gene subsets. However, growing data complexity is pushing the limits of these techniques. A review of the recent literature shows that arranging elements by similarity in image-form for a convolutional neural network (CNN) improves classification performance over treating them individually. Expanding on this, here we show a pipeline, DeepInsight-FS, to uncover gene subsets of clinical relevance. DeepInsight-FS converts non-image samples into image-form and performs element selection via CNN. To our knowledge, this is the first approach to employ CNN for element or gene selection on non-image data. A real world application of DeepInsight-FS to publicly available cancer data identified gene sets with significant overlap to several cancer-associated pathways suggesting the potential of this method to discover biomedically meaningful connections.

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

Link to bioRxiv paper:

http://biorxiv.org/cgi/content/short/2020.09.17.301168v1?rss=1

Authors: Tageldin, A., Adly, D., Mostafa, H., Mohammed, H.

Abstract:

The use of technology in agriculture has grown in recent years with the era of data analytics affecting every industry. The main challenge in using technology in agriculture is identification of effectiveness of big data analytics algorithms and their application methods. Pest management is one of the most important problems facing farmers. The cotton leafworm, Spodoptera littoralis (Boisd.) (CLW) is one of the major polyphagous key pests attacking plants includes 73 species recorded at Egypt. In the present study, several machine learning algorithms have been implemented to predict plant infestation with CLW. The moth of CLW data was weekly collected for two years in a commercial hydroponic greenhouse. Furthermore, among other features temperature and relative humidity were recorded over the total period of the study. It was proven that the XGBoost algorithm is the most effective algorithm applied in this study. Prediction accuracy of 84% has been achieved using this algorithm. The impact of environmental features on the prediction accuracy was compared with each other to ensure a complete dataset for future results. In conclusion, the present study provided a framework for applying machine learning in the prediction of plant infestation with the CLW in the greenhouses. Based on this framework, further studies with continuous measurements are warranted to achieve greater accuracy.

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