Improving deep learning method for biomedical named entity recognition by using entity definition information

Background

For differential abundance analysis, zero-inflated generalized linear models, typically zero-inflated NB models, have been increasingly used to model microbiome and other sequencing count data. A common assumption in estimating the false discovery rate is that the p values are uniformly distributed under the null hypothesis, which demands that the postulated model fit the count data adequately. Mis-specification of the distribution of the count data may lead to excess false discoveries. Therefore, model checking is critical to control the FDR at a nominal level in differential abundance analysis. Increasing studies show that the method of randomized quantile residual (RQR) performs well in diagnosing count regression models. However, the performance of RQR in diagnosing zero-inflated GLMMs for sequencing count data has not been extensively investigated in the literature.

Results

We conduct large-scale simulation studies to investigate the performance of the RQRs for zero-inflated GLMMs. The simulation studies show that the type I error rates of the GOF tests with RQRs are very close to the nominal level; in addition, the scatter-plots and Q–Q plots of RQRs are useful in discerning the good and bad models. We also apply the RQRs to diagnose six GLMMs to a real microbiome dataset. The results show that the OTU counts at the genus level of this dataset (after a truncation treatment) can be modelled well by zero-inflated and zero-modified NB models.

Conclusion

RQR is an excellent tool for diagnosing GLMMs for zero-inflated count data, particularly the sequencing count data arising in microbiome studies. In the supplementary materials, we provided two generic R functions, called rqr.glmmtmb and rqr.hurdle.glmmtmb, for calculating the RQRs given fitting outputs of the R package glmmTMB.

     

Nebula: ultra-efficient mapping-free structural variant genotyper

Large scale catalogs of common genetic variants (including indels and structural variants) are being created using data from second and third generation whole-genome sequencing technologies. However, the genotyping of these variants in newly sequenced samples is a nontrivial task that requires extensive computational resources. Furthermore, current approaches are mostly limited to only specific types of variants and are generally prone to various errors and ambiguities when genotyping complex events. We are proposing an ultra-efficient approach for genotyping any type of structural variation that is not limited by the shortcomings and complexities of current mapping-based approaches. Our method Nebula utilizes the changes in the count of k-mers to predict the genotype of structural variants. We have shown that not only Nebula is an order of magnitude faster than mapping based approaches for genotyping structural variants, but also has comparable accuracy to state-of-the-art approaches. Furthermore, Nebula is a generic framework not limited to any specific type of event. Nebula is publicly available at https://github.com/Parsoa/Nebula.     

Microbial Interaction Network Estimation via Bias-Corrected Graphical Lasso

With the increasing availability of microbiome 16S data, network estimation has become a useful approach to studying the interactions between microbial taxa. Network estimation on a set of variables is frequently explored using graphical models, in which the relationship between two variables is modeled via their conditional dependency given the other variables. Various methods for sparse inverse covariance estimation have been proposed to estimate graphical models in the high-dimensional setting, including graphical lasso. However, current methods do not address the compositional count nature of microbiome data, where abundances of microbial taxa are not directly measured, but are reflected by the observed counts in an error-prone manner. Adding to the challenge is that the sum of the counts within each sample, termed “sequencing depth,” is an experimental technicality that carries no biological information but can vary drastically across samples. To address these issues, we develop a new approach to network estimation, called BC-GLASSO (bias-corrected graphical lasso), which models the microbiome data using a logistic normal multinomial distribution with the sequencing depths explicitly incorporated, corrects the bias of the naive empirical covariance estimator arising from the heterogeneity in sequencing depths, and builds the inverse covariance estimator via graphical lasso. We demonstrate the advantage of BC-GLASSO over current approaches to microbial interaction network estimation under a variety of simulation scenarios. We also illustrate the efficacy of our method in an application to a human microbiome data set.

     

PanClassif: Improving pan cancer classification of single cell RNA-seq gene expression data using machine learning

Cancer is one of the major causes of human death per year. In recent years, cancer identification and classification using machine learning have gained momentum due to the availability of high throughput sequencing data. Using RNA-seq, cancer research is blooming day by day and new insights of cancer and related treatments are coming into light. In this paper, we propose PanClassif, a method that requires a very few and effective genes to detect cancer from RNA-seq data and is able to provide performance gain in several wide range machine learning classifiers. We have taken 22 types of cancer samples from The Cancer Genome Atlas (TCGA) having 8287 cancer samples and 680 normal samples. Firstly, PanClassif uses k-Nearest Neighbour (k-NN) smoothing to smooth the samples to handle noise in the data. Then effective genes are selected by Anova based test. For balancing the train data, PanClassif applies an oversampling method, SMOTE. We have performed comprehensive experiments on the datasets using several classification algorithms. Experimental results shows that PanClassif outperform existing state-of-the-art methods available and shows consistent performance for two single cell RNA-seq datasets taken from Gene Expression Omnibus (GEO). PanClassif improves performances of a wide variety of classifiers for both binary cancer prediction and multi-class cancer classification. PanClassif is available as a python package (https://pypi.org/project/panclassif/). All the source code and materials of PanClassif are available at https://github.com/Zwei-inc/panclassif.     

Enhanced Regulatory Sequence Prediction Using Gapped k-mer Features

Oligomers of length k, or k-mers, are convenient and widely used features for modeling the properties and functions of DNA and protein sequences. However, k-mers suffer from the inherent limitation that if the parameter k is increased to resolve longer features, the probability of observing any specific k-mer becomes very small, and k-mer counts approach a binary variable, with most k-mers absent and a few present once. Thus, any statistical learning approach using k-mers as features becomes susceptible to noisy training set k-mer frequencies once k becomes large. To address this problem, we introduce alternative feature sets using gapped k-mers, a new classifier, gkm-SVM, and a general method for robust estimation of k-mer frequencies. To make the method applicable to large-scale genome wide applications, we develop an efficient tree data structure for computing the kernel matrix. We show that compared to our original kmer-SVM and alternative approaches, our gkm-SVM predicts functional genomic regulatory elements and tissue specific enhancers with significantly improved accuracy, increasing the precision by up to a factor of two. We then show that gkm-SVM consistently outperforms kmer-SVM on human ENCODE ChIP-seq datasets, and further demonstrate the general utility of our method using a Naïve-Bayes classifier. Although developed for regulatory sequence analysis, these methods can be applied to any sequence classification problem.     

A benchmark and an algorithm for detecting germline transposon insertions and measuring de novo transposon insertion frequencies

Transposons are genomic parasites, and their new insertions can cause instability and spur the evolution of their host genomes. Rapid accumulation of short-read whole-genome sequencing data provides a great opportunity for studying new transposon insertions and their impacts on the host genome. Although many algorithms are available for detecting transposon insertions, the task remains challenging and existing tools are not designed for identifying de novo insertions. Here, we present a new benchmark fly dataset based on PacBio long-read sequencing and a new method TEMP2 for detecting germline insertions and measuring de novo ‘singleton’ insertion frequencies in eukaryotic genomes. TEMP2 achieves high sensitivity and precision for detecting germline insertions when compared with existing tools using both simulated data in fly and experimental data in fly and human. Furthermore, TEMP2 can accurately assess the frequencies of de novo transposon insertions even with high levels of chimeric reads in simulated datasets; such chimeric reads often occur during the construction of short-read sequencing libraries. By applying TEMP2 to published data on hybrid dysgenic flies inflicted by de-repressed P-elements, we confirmed the continuous new insertions of P-elements in dysgenic offspring before they regain piRNAs for P-element repression. TEMP2 is freely available at Github: https://github.com/weng-lab/TEMP2.     

coupleCoC+: An information-theoretic co-clustering-based transfer learning framework for the integrative analysis of single-cell genomic data

Technological advances have enabled us to profile multiple molecular layers at unprecedented single-cell resolution and the available datasets from multiple samples or domains are growing. These datasets, including scRNA-seq data, scATAC-seq data and sc-methylation data, usually have different powers in identifying the unknown cell types through clustering. So, methods that integrate multiple datasets can potentially lead to a better clustering performance. Here we propose coupleCoC+ for the integrative analysis of single-cell genomic data. coupleCoC+ is a transfer learning method based on the information-theoretic co-clustering framework. In coupleCoC+, we utilize the information in one dataset, the source data, to facilitate the analysis of another dataset, the target data. coupleCoC+ uses the linked features in the two datasets for effective knowledge transfer, and it also uses the information of the features in the target data that are unlinked with the source data. In addition, coupleCoC+ matches similar cell types across the source data and the target data. By applying coupleCoC+ to the integrative clustering of mouse cortex scATAC-seq data and scRNA-seq data, mouse and human scRNA-seq data, mouse cortex sc-methylation and scRNA-seq data, and human blood dendritic cells scRNA-seq data from two batches, we demonstrate that coupleCoC+ improves the overall clustering performance and matches the cell subpopulations across multimodal single-cell genomic datasets. coupleCoC+ has fast convergence and it is computationally efficient. The software is available at https://github.com/cuhklinlab/coupleCoC_plus.     

cola: an R/Bioconductor package for consensus partitioning through a general framework

Classification of high-throughput genomic data is a powerful method to assign samples to subgroups with specific molecular profiles. Consensus partitioning is the most widely applied approach to reveal subgroups by summarizing a consensus classification from a list of individual classifications generated by repeatedly executing clustering on random subsets of the data. It is able to evaluate the stability of the classification. We implemented a new R/Bioconductor package, cola, that provides a general framework for consensus partitioning. With cola, various parameters and methods can be user-defined and easily integrated into different steps of an analysis, e.g., feature selection, sample classification or defining signatures. cola provides a new method named ATC (ability to correlate to other rows) to extract features and recommends spherical k-means clustering (skmeans) for subgroup classification. We show that ATC and skmeans have better performance than other commonly used methods by a comprehensive benchmark on public datasets. We also benchmark key parameters in the consensus partitioning procedure, which helps users to select optimal parameter values. Moreover, cola provides rich functionalities to apply multiple partitioning methods in parallel and directly compare their results, as well as rich visualizations. cola can automate the complete analysis and generates a comprehensive HTML report.     

A Consistency-Based Feature Selection Method Allied with Linear SVMs for HIV-1 Protease Cleavage Site Prediction

Background

Predicting type-1 Human Immunodeficiency Virus (HIV-1) protease cleavage site in protein molecules and determining its specificity is an important task which has attracted considerable attention in the research community. Achievements in this area are expected to result in effective drug design (especially for HIV-1 protease inhibitors) against this life-threatening virus. However, some drawbacks (like the shortage of the available training data and the high dimensionality of the feature space) turn this task into a difficult classification problem. Thus, various machine learning techniques, and specifically several classification methods have been proposed in order to increase the accuracy of the classification model. In addition, for several classification problems, which are characterized by having few samples and many features, selecting the most relevant features is a major factor for increasing classification accuracy.

Results

We propose for HIV-1 data a consistency-based feature selection approach in conjunction with recursive feature elimination of support vector machines (SVMs). We used various classifiers for evaluating the results obtained from the feature selection process. We further demonstrated the effectiveness of our proposed method by comparing it with a state-of-the-art feature selection method applied on HIV-1 data, and we evaluated the reported results based on attributes which have been selected from different combinations.

Conclusion

Applying feature selection on training data before realizing the classification task seems to be a reasonable data-mining process when working with types of data similar to HIV-1. On HIV-1 data, some feature selection or extraction operations in conjunction with different classifiers have been tested and noteworthy outcomes have been reported. These facts motivate for the work presented in this paper.

Software availability

The software is available at http://ozyer.etu.edu.tr/c-fs-svm.rar.The software can be downloaded at esnag.etu.edu.tr/software/hiv_cleavage_site_prediction.rar; you will find a readme file which explains how to set the software in order to work.     

Assessing the accuracy of plotless density estimators using census counts to refine population estimates of the vultures of Kruger National Park

Breeding population estimates for three vulture species in Kruger National Park (KNP), South Africa, were made in 2013 using data from aerial censuses and a plotless density estimator (PDE). PDEs are distance-based methods used to assess sparse populations unsuitable for plot-based methods. A correction factor was applied to the 2013 estimates to reflect the difference between the survey counts and the PDE figures. We flew additional censuses across most of KNP and counted all visible nests to assess the 2013 estimates. Survey counts were within 95% confidence limits of corrected PDE estimates for White-backed Vulture Gyps africanus (count: 892; estimate: 904 [95% CI ±162]), at the limit of confidence for White-headed Vulture Trigonoceps occipitalis (count: 48; estimate: 60 [±13]) and outside confidence limits for Lappet-faced Vulture Torgos tracheliotos (count: 44; estimate: 78 [±18]). Uncorrected PDE estimates accurately reflected White-headed and Lappet-faced Vulture nest counts. The clustered patterns of White-backed Vulture nests and dispersed patterns of White-headed and Lappet-faced Vulture nests offer an explanation for these results and means that corrected PDE densities are inaccurate for estimating dispersed nests but accurate for estimating clustered nests. Using PDE methods, aerial surveys over ～35% of KNP are probably sufficient to assess changes in these vulture populations over time. Our results highlight these globally important breeding populations.     

Heterogeneous Ensemble Combination Search Using Genetic Algorithm for Class Imbalanced Data Classification

Classification of datasets with imbalanced sample distributions has always been a challenge. In general, a popular approach for enhancing classification performance is the construction of an ensemble of classifiers. However, the performance of an ensemble is dependent on the choice of constituent base classifiers. Therefore, we propose a genetic algorithm-based search method for finding the optimum combination from a pool of base classifiers to form a heterogeneous ensemble. The algorithm, called GA-EoC, utilises 10 fold-cross validation on training data for evaluating the quality of each candidate ensembles. In order to combine the base classifiers decision into ensemble’s output, we used the simple and widely used majority voting approach. The proposed algorithm, along with the random sub-sampling approach to balance the class distribution, has been used for classifying class-imbalanced datasets. Additionally, if a feature set was not available, we used the (α, β) − k Feature Set method to select a better subset of features for classification. We have tested GA-EoC with three benchmarking datasets from the UCI-Machine Learning repository, one Alzheimer’s disease dataset and a subset of the PubFig database of Columbia University. In general, the performance of the proposed method on the chosen datasets is robust and better than that of the constituent base classifiers and many other well-known ensembles. Based on our empirical study we claim that a genetic algorithm is a superior and reliable approach to heterogeneous ensemble construction and we expect that the proposed GA-EoC would perform consistently in other cases.     

CoMeta: Classification of Metagenomes Using k-mers

Nowadays, the study of environmental samples has been developing rapidly. Characterization of the environment composition broadens the knowledge about the relationship between species composition and environmental conditions. An important element of extracting the knowledge of the sample composition is to compare the extracted fragments of DNA with sequences derived from known organisms. In the presented paper, we introduce an algorithm called CoMeta (Classification of metagenomes), which assigns a query read (a DNA fragment) into one of the groups previously prepared by the user. Typically, this is one of the taxonomic rank (e.g., phylum, genus), however prepared groups may contain sequences having various functions. In CoMeta, we used the exact method for read classification using short subsequences (k-mers) and fast program for indexing large set of k-mers. In contrast to the most popular methods based on BLAST, where the query is compared with each reference sequence, we begin the classification from the top of the taxonomy tree to reduce the number of comparisons. The presented experimental study confirms that CoMeta outperforms other programs used in this context. CoMeta is available at https://github.com/jkawulok/cometa under a free GNU GPL 2 license.     

On the statistical assessment of classifiers using DNA microarray data

Background

In this paper we present a method for the statistical assessment of cancer predictors which make use of gene expression profiles. The methodology is applied to a new data set of microarray gene expression data collected in Casa Sollievo della Sofferenza Hospital, Foggia – Italy. The data set is made up of normal (22) and tumor (25) specimens extracted from 25 patients affected by colon cancer. We propose to give answers to some questions which are relevant for the automatic diagnosis of cancer such as: Is the size of the available data set sufficient to build accurate classifiers? What is the statistical significance of the associated error rates? In what ways can accuracy be considered dependant on the adopted classification scheme? How many genes are correlated with the pathology and how many are sufficient for an accurate colon cancer classification? The method we propose answers these questions whilst avoiding the potential pitfalls hidden in the analysis and interpretation of microarray data.

Results

We estimate the generalization error, evaluated through the Leave-K-Out Cross Validation error, for three different classification schemes by varying the number of training examples and the number of the genes used. The statistical significance of the error rate is measured by using a permutation test. We provide a statistical analysis in terms of the frequencies of the genes involved in the classification. Using the whole set of genes, we found that the Weighted Voting Algorithm (WVA) classifier learns the distinction between normal and tumor specimens with 25 training examples, providing e = 21% (p = 0.045) as an error rate. This remains constant even when the number of examples increases. Moreover, Regularized Least Squares (RLS) and Support Vector Machines (SVM) classifiers can learn with only 15 training examples, with an error rate of e = 19% (p = 0.035) and e = 18% (p = 0.037) respectively. Moreover, the error rate decreases as the training set size increases, reaching its best performances with 35 training examples. In this case, RLS and SVM have error rates of e = 14% (p = 0.027) and e = 11% (p = 0.019). Concerning the number of genes, we found about 6000 genes (p < 0.05) correlated with the pathology, resulting from the signal-to-noise statistic. Moreover the performances of RLS and SVM classifiers do not change when 74% of genes is used. They progressively reduce up to e = 16% (p < 0.05) when only 2 genes are employed. The biological relevance of a set of genes determined by our statistical analysis and the major roles they play in colorectal tumorigenesis is discussed.

Conclusions

The method proposed provides statistically significant answers to precise questions relevant for the diagnosis and prognosis of cancer. We found that, with as few as 15 examples, it is possible to train statistically significant classifiers for colon cancer diagnosis. As for the definition of the number of genes sufficient for a reliable classification of colon cancer, our results suggest that it depends on the accuracy required.     

Waste Not,Want Not: Why Rarefying Microbiome Data Is Inadmissible

Current practice in the normalization of microbiome count data is inefficient in the statistical sense. For apparently historical reasons, the common approach is either to use simple proportions (which does not address heteroscedasticity) or to use rarefying of counts, even though both of these approaches are inappropriate for detection of differentially abundant species. Well-established statistical theory is available that simultaneously accounts for library size differences and biological variability using an appropriate mixture model. Moreover, specific implementations for DNA sequencing read count data (based on a Negative Binomial model for instance) are already available in RNA-Seq focused R packages such as edgeR and DESeq. Here we summarize the supporting statistical theory and use simulations and empirical data to demonstrate substantial improvements provided by a relevant mixture model framework over simple proportions or rarefying. We show how both proportions and rarefied counts result in a high rate of false positives in tests for species that are differentially abundant across sample classes. Regarding microbiome sample-wise clustering, we also show that the rarefying procedure often discards samples that can be accurately clustered by alternative methods. We further compare different Negative Binomial methods with a recently-described zero-inflated Gaussian mixture, implemented in a package called metagenomeSeq. We find that metagenomeSeq performs well when there is an adequate number of biological replicates, but it nevertheless tends toward a higher false positive rate. Based on these results and well-established statistical theory, we advocate that investigators avoid rarefying altogether. We have provided microbiome-specific extensions to these tools in the R package, phyloseq.     

MEGA11: Molecular Evolutionary Genetics Analysis Version 11

The Molecular Evolutionary Genetics Analysis (MEGA) software has matured to contain a large collection of methods and tools of computational molecular evolution. Here, we describe new additions that make MEGA a more comprehensive tool for building timetrees of species, pathogens, and gene families using rapid relaxed-clock methods. Methods for estimating divergence times and confidence intervals are implemented to use probability densities for calibration constraints for node-dating and sequence sampling dates for tip-dating analyses. They are supported by new options for tagging sequences with spatiotemporal sampling information, an expanded interactive Node Calibrations Editor, and an extended Tree Explorer to display timetrees. Also added is a Bayesian method for estimating neutral evolutionary probabilities of alleles in a species using multispecies sequence alignments and a machine learning method to test for the autocorrelation of evolutionary rates in phylogenies. The computer memory requirements for the maximum likelihood analysis are reduced significantly through reprogramming, and the graphical user interface has been made more responsive and interactive for very big data sets. These enhancements will improve the user experience, quality of results, and the pace of biological discovery. Natively compiled graphical user interface and command-line versions of MEGA11 are available for Microsoft Windows, Linux, and macOS from www.megasoftware.net.     

Inferring Adaptive Introgression Using Hidden Markov Models

Adaptive introgression—the flow of adaptive genetic variation between species or populations—has attracted significant interest in recent years and it has been implicated in a number of cases of adaptation, from pesticide resistance and immunity, to local adaptation. Despite this, methods for identification of adaptive introgression from population genomic data are lacking. Here, we present Ancestry_HMM-S, a hidden Markov model-based method for identifying genes undergoing adaptive introgression and quantifying the strength of selection acting on them. Through extensive validation, we show that this method performs well on moderately sized data sets for realistic population and selection parameters. We apply Ancestry_HMM-S to a data set of an admixed Drosophila melanogaster population from South Africa and we identify 17 loci which show signatures of adaptive introgression, four of which have previously been shown to confer resistance to insecticides. Ancestry_HMM-S provides a powerful method for inferring adaptive introgression in data sets that are typically collected when studying admixed populations. This method will enable powerful insights into the genetic consequences of admixture across diverse populations. Ancestry_HMM-S can be downloaded from https://github.com/jesvedberg/Ancestry_HMM-S/.