Dataset Information

Identification of important regressor groups, subgroups and individuals via regularization methods: application to gut microbiome data.

ABSTRACT:

Motivation

Gut microbiota can be classified at multiple taxonomy levels. Strategies to use changes in microbiota composition to effect health improvements require knowing at which taxonomy level interventions should be aimed. Identifying these important levels is difficult, however, because most statistical methods only consider when the microbiota are classified at one taxonomy level, not multiple.

Results

Using L1 and L2 regularizations, we developed a new variable selection method that identifies important features at multiple taxonomy levels. The regularization parameters are chosen by a new, data-adaptive, repeated cross-validation approach, which performed well. In simulation studies, our method outperformed competing methods: it more often selected significant variables, and had small false discovery rates and acceptable false-positive rates. Applying our method to gut microbiota data, we found which taxonomic levels were most altered by specific interventions or physiological status.

Availability

The new approach is implemented in an R package, which is freely available from the corresponding author.

Contact

tpgarcia@srph.tamhsc.edu

Supplementary information

Supplementary data are available at Bioinformatics online.

SUBMITTER: Garcia TP

PROVIDER: S-EPMC3957069 | biostudies-literature | 2014 Mar

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Identification of important regressor groups, subgroups and individuals via regularization methods: application to gut microbiome data.

Garcia Tanya P TP Müller Samuel S Carroll Raymond J RJ Walzem Rosemary L RL

Bioinformatics (Oxford, England) 20131024 6

<h4>Motivation</h4>Gut microbiota can be classified at multiple taxonomy levels. Strategies to use changes in microbiota composition to effect health improvements require knowing at which taxonomy level interventions should be aimed. Identifying these important levels is difficult, however, because most statistical methods only consider when the microbiota are classified at one taxonomy level, not multiple.<h4>Results</h4>Using L1 and L2 regularizations, we developed a new variable selection met ...[more]

PMID: 24162467

Similar Datasets

Project description:Objective: The gut microecosystem is the largest microecosystem in the human body and has been proven to be linked to neurological diseases. The main objective of this study was to characterize the fecal microbiome, investigate the differences between epilepsy patients and healthy controls, and evaluate the potential efficacy of the fecal microbiome as a diagnostic tool for epilepsy. Design: We collected 74 fecal samples from epilepsy patients (Eps, n = 24) and healthy controls (HCs, n = 50) in the First Affiliated Hospital of Zhengzhou University and subjected the samples to 16S rRNA MiSeq sequencing and analysis. We set up a train set and a test set, identified the optimal microbial markers for epilepsy after characterizing the gut microbiome in the former and built a diagnostic model, then validated it in the validation group. Results: There were significant differences in microbial communities between the two groups. The α-diversity of the HCs was higher than that of the epilepsy group, but the Venn diagram showed that there were more unique operational taxonomic unit (OTU) in the epilepsy group. At the phylum level, Proteobacteria and Actinobacteriota increased significantly in Eps, while the relative abundance of Bacteroidota increased in HCs. Compared with HCs, Eps were enriched in 23 genera, including Faecalibacterium, Escherichia-Shigella, Subdoligranulum and Enterobacteriaceae-unclassified. In contrast, 59 genera including Bacteroides, Megamonas, Prevotella, Lachnospiraceae-unclassified and Blautia increased in the HCs. In Spearman correlation analysis, age, WBC, RBC, PLT, ALB, CREA, TBIL, Hb and Urea were positively correlated with most of the different OTUs. Seizure-type, course and frequency are negatively correlated with most of the different OTUs. In addition, twenty-two optimal microbial markers were identified by a fivefold cross-validation of the random forest model. In the established train set and test set, the area under the curve was 0.9771 and 0.993, respectively. Conclusion: Our study was the first to characterize the gut microbiome of Eps and HCs in central China and demonstrate the potential efficacy of microbial markers as a noninvasive biological diagnostic tool for epilepsy.

Project description:Background: While CRISPR-Cas systems have been identified in bacteria from a wide variety of ecological niches, there are no studies to describe CRISPR-Cas elements in Bacteroides species, the most prevalent anaerobic bacteria in the lower intestinal tract. Microbes of the genus Bacteroides make up ~25% of the total gut microbiome. Bacteroides fragilis comprises only 2% of the total Bacteroides in the gut, yet causes of >70% of Bacteroides infections. The factors causing it to transition from benign resident of the gut microbiome to virulent pathogen are not well understood, but a combination of horizontal gene transfer (HGT) of virulence genes and differential transcription of endogenous genes are clearly involved. The CRISPR-Cas system is a multi-functional system described in prokaryotes that may be involved in control both of HGT and of gene regulation. Results: Clustered regularly interspaced short palindromic repeats (CRISPR) elements in all strains of B. fragilis (n = 109) with publically available genomes were identified. Three different CRISPR-Cas types, corresponding most closely to Type IB, Type IIIB, and Type IIC, were identified. Thirty-five strains had two CRISPR-Cas types, and three strains included all three CRISPR-Cas types in their respective genomes. The cas1 gene in the Type IIIB system encoded a reverse-transcriptase/Cas1 fusion protein rarely found in prokaryotes. We identified a short CRISPR (3 DR) with no associated cas genes present in most of the isolates; these CRISPRs were found immediately upstream of a hipA/hipB operon and we speculate that this element may be involved in regulation of this operon related to formation of persister cells during antimicrobial exposure. Also, blood isolates of B. fragilis did not have Type IIC CRISPR-Cas systems and had atypical Type IIIB CRISPR-Cas systems that were lacking adjacent cas genes. Conclusions: This is the first systematic report of CRISPR-Cas systems in a wide range of B. fragilis strains from a variety of sources. There are four apparent CRISPR-Cas systems in B. fragilis-three systems have adjacent cas genes. Understanding CRISPR/Cas function in B. fragilis will elucidate their role in gene expression, DNA repair and ability to survive exposure to antibiotics. Also, based on their unique CRISPR-Cas arrays, their phylogenetic clustering and their virulence potential, we are proposing that blood isolates of B. fragilis be viewed a separate subgroup.

Dataset Information

Identification of important regressor groups, subgroups and individuals via regularization methods: application to gut microbiome data.

Motivation

Results

Availability

Contact

Supplementary information

Publications

Identification of important regressor groups, subgroups and individuals via regularization methods: application to gut microbiome data.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets