Rob D. Knight

TEMPTED improves analysis of longitudinal microbiome data by incorporating time as a continuous variable, achieving high accuracy in phenotype classification and identifying microbial markers linked to birth outcomes.

The paper creates a spectral library of N-acyl lipids from metabolomics data, identifies novel compounds, and links them to health conditions like HIV and cognitive impairment.

The paper develops a drug exposure resource from untargeted metabolomics data to identify medications without clinical records.

The paper genotypes Y and mitochondrial chromosomes in heterogeneous stock rats and finds no significant associations with complex traits, except for gene expression on these chromosomes.

This study identifies genetic factors enabling Bacteroides fragilis to transition from commensal to pathogenic by analyzing 813 strains and their genomic diversity.

The paper introduces a weighted variance components framework to integrate metagenomic and metabolomic data for pathway analysis, showing improved power while controlling type I error.

This study compares the stability of fecal metabolome and microbiome using 95% ethanol and commercial kits under different storage conditions, finding comparable results to flash freezing.

This study integrates polygenic risk scores and gut metagenomic data to improve prediction of common diseases, showing enhanced predictive performance over conventional risk factors.

This study compares the pangenome of Bacteroides fragilis genomospecies to reveal genetic diversity and ecological roles, finding division I and II strains represent distinct species with different functional adaptations.

This paper engineers bacteria to detect tumor DNA through horizontal gene transfer, achieving 100% discrimination between mice with and without colorectal cancer.

This paper introduces a deep learning method that detects multiple diseases simultaneously using microbiome data and host variables, outperforming traditional ML approaches and capturing key disease-related factors.

This paper introduces C-DEPP, a method that scales deep phylogenetic embedding to large reference trees by using an ensemble approach, enabling efficient placement of 20 million 16S fragments on the GG2 tree in 41 hours.

The authors reanalyze their pan-cancer microbiome findings using decontaminated data to address concerns about contamination, batch effects, and machine learning approaches.

BIRDMAn is a Bayesian framework for differential abundance analysis in microbiome studies, showing improved statistical power and detecting antibiotic effects and cancer classifiers.

Greengenes2 aligns 16S rRNA and shotgun metagenomic data using a whole-genome phylogeny, showing agreement in principal coordinates, taxonomy, and phenotype effect size.

The paper shows that sample collection time significantly affects microbiome study conclusions, more than experimental interventions or dietary changes.

This study uses native E. coli as chassis to deliver transgenes, inducing persistent physiological changes in conventionally-raised hosts.

Tensor-cell2cell improves cell-cell communication analysis by incorporating cellular context, identifying communication modules linked to diseases like COVID-19 and Autism Spectrum Disorder.

This paper tests cleaning solutions for sentinel cards to improve temporal resolution of SARS-CoV-2 monitoring in schools by distinguishing new infections from persistent viral traces.

Evident calculates effect sizes and performs power analysis for microbiome community differences, scaling to large datasets with many metadata covariates.

This study uses gut microbiome data and machine learning to improve liver disease risk prediction, showing that combining microbiome features with conventional risk factors enhances model performance.

The paper predicts structures and annotates functions for diverse microbial protein sequences, identifying novel folds and highlighting the saturation of structural space.

In 5,959 Finns with genotypes and stool metagenomes, GWAS found 583 independent SNP-taxon associations; dairy intake modified LCT effects on Bifidobacterium, and ABO-linked Faecalicatena lactaris tracked secretor status.

The paper introduces a kernel-based method to identify host genetic variants associated with microbiome beta-diversity, finding immunity-related genes influence gut microbiota composition.

This paper introduces a Graph Convolutional Network for predicting protein function and identifying important residues, showing improved accuracy over sequence-based methods.

This paper highlights challenges in benchmarking metagenomic profilers due to confusion between sequence abundance and taxonomic abundance.

This paper introduces OGUs, a metagenome analysis method that uses sequence alignment hits to reference genomes for community analysis, showing better performance in predicting biological traits than taxonomic or 16S rRNA methods.

EMPress provides an interactive tool for visualizing phylogenetic trees alongside multi-omic data, enabling exploratory analyses of microbiome and metabolome datasets.

The paper identifies 31 host genetic loci influencing gut microbiome composition through GWAS and suggests causal effects in inflammatory diseases.

Dual patterned unique molecular identifiers (UMIs) turn Nanopore or PacBio CCS reads of ~4400 bp rRNA operons into single-molecule consensus sequences with 0.0049% (ONT) or 0.0006% (PacBio) mean error.