Dataset Information

Machine Learning Meta-analysis of Large Metagenomic Datasets: Tools and Biological Insights.

ABSTRACT: Shotgun metagenomic analysis of the human associated microbiome provides a rich set of microbial features for prediction and biomarker discovery in the context of human diseases and health conditions. However, the use of such high-resolution microbial features presents new challenges, and validated computational tools for learning tasks are lacking. Moreover, classification rules have scarcely been validated in independent studies, posing questions about the generality and generalization of disease-predictive models across cohorts. In this paper, we comprehensively assess approaches to metagenomics-based prediction tasks and for quantitative assessment of the strength of potential microbiome-phenotype associations. We develop a computational framework for prediction tasks using quantitative microbiome profiles, including species-level relative abundances and presence of strain-specific markers. A comprehensive meta-analysis, with particular emphasis on generalization across cohorts, was performed in a collection of 2424 publicly available metagenomic samples from eight large-scale studies. Cross-validation revealed good disease-prediction capabilities, which were in general improved by feature selection and use of strain-specific markers instead of species-level taxonomic abundance. In cross-study analysis, models transferred between studies were in some cases less accurate than models tested by within-study cross-validation. Interestingly, the addition of healthy (control) samples from other studies to training sets improved disease prediction capabilities. Some microbial species (most notably Streptococcus anginosus) seem to characterize general dysbiotic states of the microbiome rather than connections with a specific disease. Our results in modelling features of the "healthy" microbiome can be considered a first step toward defining general microbial dysbiosis. The software framework, microbiome profiles, and metadata for thousands of samples are publicly available at http://segatalab.cibio.unitn.it/tools/metaml.

SUBMITTER: Pasolli E

PROVIDER: S-EPMC4939962 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundMetagenomics is an approach to the characterization of microbial genomes via the direct isolation of genomic sequences from the environment without prior cultivation. The amount of metagenomic sequence data is growing fast while computational methods for metagenome analysis are still in their infancy. In contrast to genomic sequences of single species, which can usually be assembled and analyzed by many available methods, a large proportion of metagenome data remains as unassembled anonymous sequencing reads. One of the aims of all metagenomic sequencing projects is the identification of novel genes. Short length, for example, Sanger sequencing yields on average 700 bp fragments, and unknown phylogenetic origin of most fragments require approaches to gene prediction that are different from the currently available methods for genomes of single species. In particular, the large size of metagenomic samples requires fast and accurate methods with small numbers of false positive predictions.ResultsWe introduce a novel gene prediction algorithm for metagenomic fragments based on a two-stage machine learning approach. In the first stage, we use linear discriminants for monocodon usage, dicodon usage and translation initiation sites to extract features from DNA sequences. In the second stage, an artificial neural network combines these features with open reading frame length and fragment GC-content to compute the probability that this open reading frame encodes a protein. This probability is used for the classification and scoring of gene candidates. With large scale training, our method provides fast single fragment predictions with good sensitivity and specificity on artificially fragmented genomic DNA. Additionally, this method is able to predict translation initiation sites accurately and distinguishes complete from incomplete genes with high reliability.ConclusionLarge scale machine learning methods are well-suited for gene prediction in metagenomic DNA fragments. In particular, the combination of linear discriminants and neural networks is promising and should be considered for integration into metagenomic analysis pipelines. The data sets can be downloaded from the URL provided (see Availability and requirements section).

Project description:BackgroundShotgun metagenomics based on untargeted sequencing can explore the taxonomic profile and the function of unknown microorganisms in samples, and complement the shortage of amplicon sequencing. Binning assembled sequences into individual groups, which represent microbial genomes, is the key step and a major challenge in metagenomic research. Both supervised and unsupervised machine learning methods have been employed in binning. Genome binning belonging to unsupervised method clusters contigs into individual genome bins by machine learning methods without the assistance of any reference databases. So far a lot of genome binning tools have emerged. Evaluating these genome tools is of great significance to microbiological research. In this study, we evaluate 15 genome binning tools containing 12 original binning tools and 3 refining binning tools by comparing the performance of these tools on chicken gut metagenomic datasets and the first CAMI challenge datasets.ResultsFor chicken gut metagenomic datasets, original genome binner MetaBat, Groopm2 and Autometa performed better than other original binner, and MetaWrap combined the binning results of them generated the most high-quality genome bins. For CAMI datasets, Groopm2 achieved the highest purity (> 0.9) with good completeness (> 0.8), and reconstructed the most high-quality genome bins among original genome binners. Compared with Groopm2, MetaBat2 had similar performance with higher completeness and lower purity. Genome refining binners DASTool predicated the most high-quality genome bins among all genomes binners. Most genome binner performed well for unique strains. Nonetheless, reconstructing common strains still is a substantial challenge for all genome binner.ConclusionsIn conclusion, we tested a set of currently available, state-of-the-art metagenomics hybrid binning tools and provided a guide for selecting tools for metagenomic binning by comparing range of purity, completeness, adjusted rand index, and the number of high-quality reconstructed bins. Furthermore, available information for future binning strategy were concluded.

Project description:BackgroundMetagenomics is the application of modern genomic techniques to investigate the members of a microbial community directly in their natural environments and is widely used in many studies to survey the communities of microbial organisms that live in diverse ecosystems. In order to understand the metagenomic profile of one of the densest interaction spaces for millions of people, the public transit system, the MetaSUB international Consortium has collected and sequenced metagenomes from subways of different cities across the world. In collaboration with CAMDA, MetaSUB has made the metagenomic samples from these cities available for an open challenge of data analysis including, but not limited in scope to, the identification of unknown samples.ResultsTo distinguish the metagenomic profiling among different cities and also predict unknown samples precisely based on the profiling, two different approaches are proposed using machine learning techniques; one is a read-based taxonomy profiling of each sample and prediction method, and the other is a reduced representation assembly-based method. Among various machine learning techniques tested, the random forest technique showed promising results as a suitable classifier for both approaches. Random forest models developed from read-based taxonomic profiling could achieve an accuracy of 91% with 95% confidence interval between 80 and 93%. The assembly-based random forest model prediction also reached 90% accuracy. However, both models achieved roughly the same accuracy on the testing test, whereby they both failed to predict the most abundant label.ConclusionOur results suggest that both read-based and assembly-based approaches are powerful tools for the analysis of metagenomics data. Moreover, our results suggest that reduced representation assembly-based methods are able to simultaneous provide high-accuracy prediction on available data. Overall, we show that metagenomic samples can be traced back to their location with careful generation of features from the composition of microbes and utilizing existing machine learning algorithms. Proposed approaches show high accuracy of prediction, but require careful inspection before making any decisions due to sample noise or complexity.ReviewersThis article was reviewed by Eugene V. Koonin, Jing Zhou and Serghei Mangul.

Project description:OBJECTIVES:The goal of this study was to use machine learning to more accurately predict survival after echocardiography. BACKGROUND:Predicting patient outcomes (e.g., survival) following echocardiography is primarily based on ejection fraction (EF) and comorbidities. However, there may be significant predictive information within additional echocardiography-derived measurements combined with clinical electronic health record data. METHODS:Mortality was studied in 171,510 unselected patients who underwent 331,317 echocardiograms in a large regional health system. The authors investigated the predictive performance of nonlinear machine learning models compared with that of linear logistic regression models using 3 different inputs: 1) clinical variables, including 90 cardiovascular-relevant International Classification of Diseases, Tenth Revision, codes, and age, sex, height, weight, heart rate, blood pressures, low-density lipoprotein, high-density lipoprotein, and smoking; 2) clinical variables plus physician-reported EF; and 3) clinical variables and EF, plus 57 additional echocardiographic measurements. Missing data were imputed with a multivariate imputation by using a chained equations algorithm (MICE). The authors compared models versus each other and baseline clinical scoring systems by using a mean area under the curve (AUC) over 10 cross-validation folds and across 10 survival durations (6 to 60 months). RESULTS:Machine learning models achieved significantly higher prediction accuracy (all AUC >0.82) over common clinical risk scores (AUC = 0.61 to 0.79), with the nonlinear random forest models outperforming logistic regression (p < 0.01). The random forest model including all echocardiographic measurements yielded the highest prediction accuracy (p < 0.01 across all models and survival durations). Only 10 variables were needed to achieve 96% of the maximum prediction accuracy, with 6 of these variables being derived from echocardiography. Tricuspid regurgitation velocity was more predictive of survival than LVEF. In a subset of studies with complete data for the top 10 variables, multivariate imputation by chained equations yielded slightly reduced predictive accuracies (difference in AUC of 0.003) compared with the original data. CONCLUSIONS:Machine learning can fully utilize large combinations of disparate input variables to predict survival after echocardiography with superior accuracy.

Dataset Information

Machine Learning Meta-analysis of Large Metagenomic Datasets: Tools and Biological Insights.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets