Biology subjects

Korkin, D.

Publications and source records attributed to Korkin, D..

6 recordsLinked to original sources

bioRxiv2018

The genome of the soybean cyst nematode (Heterodera glycines) reveals complex patterns of duplications involved in the evolution of parasitism genes

Heterodera glycines, commonly referred to as the soybean cyst nematode (SCN), is an obligatory and sedentary plant parasite that causes over a billion-dollar yield loss to soybean production annually. Although there are genetic determinants that render soybean plants resistant to certain nematode genotypes, resistant soybean cultivars are increasingly ineffective because their multi-year usage has selected for virulent H. glycines populations. The parasitic success of H. glycines relies on the comprehensive re-engineering of an infection site into a syncytium, as well as the long-term suppression of host defense to ensure syncytial viability. At the forefront of these complex molecular interactions are effectors, the proteins secreted by H. glycines into host root tissues. The mechanisms of effector acquisition, diversification, and selection need to be understood before effective control strategies can be developed, but the lack of an annotated genome has been a major roadblock. Here, we use PacBio long-read technology to assemble a H. glycines genome of 738 contigs into 123Mb with annotations for 29,769 genes. The genome contains significant numbers of repeats (34%), tandem duplicates (18.7Mb), and horizontal gene transfer events (151 genes). Using previously published effector sequences, the newly generated H. glycines genome, and comparisons to other nematode genomes, we investigate the evolutionary mechanisms responsible for the emergence and diversification of effector genes.

genomics↗

bioRxiv2018

Multilayer view of pathogenic SNVs in human interactome through in-silico edgetic profiling

Non-synonymous mutations linked to the complex diseases often have a global impact on a biological system, affecting large biomolecular networks and pathways. However, the magnitude of the mutation-driven effects on the macromolecular network is yet to be fully explored. In this work, we present an systematic multi-level characterization of human mutations associated with genetic disorders by determining their individual and combined interaction-rewiring, \"edgetic\", effects on the human interactome. Our in-silico analysis highlights the intrinsic differences and important similarities between the pathogenic single nucleotide variants (SNVs) and frameshift mutations. We show that pathogenic SNVs are more likely to cause gene pleiotropy than pathogenic frameshift mutations and are enriched on the protein interaction interfaces. Functional profiling of SNVs indicates widespread disruption of the protein-protein interactions and synergistic effects of SNVs. The coverage of our approach is several times greater than the recently published experimental study and has the minimal overlap with it, while the distributions of determined edgotypes between the two sets of profiled mutations are remarkably similar. Case studies reveal the central role of interaction-disrupting mutations in type 2 diabetes mellitus, and suggest the importance of studying mutations that abnormally strengthen the protein interactions in cancer. With the advancement of next-generation sequencing technology that drives precision medicine, there is an increasing demand in understanding the changes in molecular mechanisms caused by the patient-specific genetic variation. The current and future in-silico edgotyping tools present a cheap and fast solution to deal with the rapidly growing datasets of discovered mutations.

systems biology↗

bioRxiv2018

Determining rewiring effects of alternatively spliced isoforms on protein-protein interactions using a computational approach

The critical role of alternative splicing (AS) in cell functioning has recently become apparent, whether in studying tissue-or cell-specific regulation, or understanding molecular mechanisms governing a complex disorder. Studying the rewiring, or edgetic, effects of alternatively spliced isoforms on protein interactome can provide system-wide insights into these questions. Unfortunately, high-throughput experiments for such studies are expensive and time-consuming, hence the need to develop an in-silico approach. Here, we formulated the problem of characterization the edgetic effects of AS on protein-protein interactions (PPIs) as a binary classification problem and introduced a first computational approach to solve it. We first developed a supervised feature-based classifier that benefited from the traditional features describing a PPI, the problem-specific features that characterized the difference between the reference and alternative isoforms, and a novel domain interaction potential that allowed pinpointing the domains employed during a specific PPI. We then expanded this approach by including a large set of unlabeled interactomics data and developing a semi-supervised learning method. Our method called AS-IN (Alternatively Splicing INteraction prediction) Tool was compared with the state-of-the-art PPI prediction tools and showed a superior performance, achieving 0.92 in precision and recall. We demonstrated the utility of AS-IN Tool by applying it to the transcriptomic data obtained from the brain and liver tissues of a healthy mouse and western diet fed mouse that developed type two diabetes. We showed that the edgetic effects of differentially expressed transcripts associated with the disease condition are system-wide and unlikely to be detected by looking only at the gene-specific expression levels.

bioinformatics↗

bioRxiv2018

Genome-wide prediction of bacterial effectors across six secretion system types using a feature-based supervised learning framework

Gram-negative bacteria are responsible for hundreds of millions infections worldwide, including the emerging hospital-acquired infections and neglected tropical diseases in the third-world countries. Finding a fast and cheap way to understand the molecular mechanisms behind the bacterial infections is critical for efficient diagnostics and treatment. An important step towards understanding these mechanisms is discovering bacterial effectors, the proteins secreted into the host through one of the six common secretion system types. Unfortunately, current effector prediction methods are designed to specifically target one of three secretion systems, and no accurate \"secretion system-agnostic\" method is available.\n\nHere, we present PREFFECTOR, a computational feature-based approach to discover effectors in Gram-negative bacteria without prior knowledge on bacterial secretion system(s) or cryptic secretion signals. Our approach was first evaluated using several assessment protocols on a manually curated, balanced dataset of experimentally determined effectors across all six secretion systems as well as non-effector proteins. The evaluation revealed high accuracy of the top performing classifiers in PREFFECTOR, with the small false positive discovery rate across all six secretion systems. Our method was also applied to four bacteria that had limited knowledge on virulence factors or secreted effectors. PREFFECTOR web-server is freely available at: http://korkinlab.org/preffector.

bioinformatics↗

bioRxiv2017

Boosting Gene Expression Clustering with System-Wide Biological Information: A Robust Autoencoder Approach

Gene expression analysis provides genome-wide insights into the transcriptional activity of a cell. One of the first computational steps in exploration and analysis of the gene expression data is clustering. With a number of standard clustering methods routinely used, most of the methods do not take prior biological information into account. In this paper, we propose a new approach for gene expression clustering analysis. The approach benefits from a new deep learning architecture, Robust Autoencoder, which provides a more accurate high-level representation of the feature sets, and from incorporating prior biological information into the clustering process. We tested our approach on two distinct gene expression datasets and compared the performance with two widely used clustering methods, hierarchical clustering and k-means, as well as with a recent deep learning clustering approach. As a result, our approach outperformed all other clustering methods on the labeled yeast gene expression dataset. Furthermore we showed that it is better in identifying the functionally common clusters than k-means on the unlabeled human gene expression dataset. The results demonstrate that our new deep learning architecture could generalize well the specific properties of gene expression profiles. Furthermore, the results confirm our hypothesis that the prior biological network knowledge could be helpful in the gene expression clustering task.

bioinformatics↗

bioRxiv2017

Biological classification with RNA-Seq data: Can alternative splicing enhance machine learning classifier?

The extent to which the genes are expressed in the cell can be simplistically defined as a function of one or more factors of the environment, lifestyle, and genetics. RNA sequencing (RNA-Seq) is becoming a prevalent approach to quantify gene expression, and is expected to gain better insights to a number of biological and biomedical questions, compared to the DNA microarrays. Most importantly, RNA-Seq allows to quantify expression at the gene and alternative splicing isoform levels. However, leveraging the RNA-Seq data requires development of new data mining and analytics methods. Supervised machine learning methods are commonly used approaches for biological data analysis, and have recently gained attention for their applications to the RNA-Seq data.\n\nIn this work, we assess the utility of supervised learning methods trained on RNA-Seq data for a diverse range of biological classification tasks. We hypothesize that the isoform-level expression data is more informative for biological classification tasks than the gene-level expression data. Our large-scale assessment is done through utilizing multiple datasets, organisms, lab groups, and RNA-Seq analysis pipelines. Overall, we performed and assessed 61 biological classification problems that leverage three independent RNA-Seq datasets and include over 2,000 samples that come from multiple organisms, lab groups, and RNA-Seq analyses. These 61 problems include predictions of the tissue type, sex, or age of the sample, healthy or cancerous phenotypes and, the pathological tumor stage for the samples from the cancerous tissue. For each classification problem, the performance of three normalization techniques and six machine learning classifiers was explored. We find that for every single classification problem, the isoform-based classifiers outperform or are comparable with gene expression based methods. The top-performing supervised learning techniques reached a near perfect classification accuracy, demonstrating the utility of supervised learning for RNA-Seq based data analysis.

bioinformatics↗