Skip to main content
eScholarship
Open Access Publications from the University of California

Scholarly Works (15 results)

  • Article
  • Peer Reviewed
UC Berkeley Previously Published Works (2018)

Motivation

Microbial communities play important roles in the function and maintenance of various biosystems, ranging from the human body to the environment. A major challenge in microbiome research is the classification of microbial communities of different environments or host phenotypes. The most common and cost-effective approach for such studies to date is 16S rRNA gene sequencing. Recent falls in sequencing costs have increased the demand for simple, efficient and accurate methods for rapid detection or diagnosis with proved applications in medicine, agriculture and forensic science. We describe a reference- and alignment-free approach for predicting environments and host phenotypes from 16S rRNA gene sequencing based on k-mer representations that benefits from a bootstrapping framework for investigating the sufficiency of shallow sub-samples. Deep learning methods as well as classical approaches were explored for predicting environments and host phenotypes.

Results

A k-mer distribution of shallow sub-samples outperformed Operational Taxonomic Unit (OTU) features in the tasks of body-site identification and Crohn's disease prediction. Aside from being more accurate, using k-mer features in shallow sub-samples allows (i) skipping computationally costly sequence alignments required in OTU-picking and (ii) provided a proof of concept for the sufficiency of shallow and short-length 16S rRNA sequencing for phenotype prediction. In addition, k-mer features predicted representative 16S rRNA gene sequences of 18 ecological environments, and 5 organismal environments with high macro-F1 scores of 0.88 and 0.87. For large datasets, deep learning outperformed classical methods such as Random Forest and Support Vector Machine.

Availability and implementation

The software and datasets are available at https://llp.berkeley.edu/micropheno.

Supplementary information

Supplementary data are available at Bioinformatics online.
    Cover page: MicroPheno: predicting environments and host phenotypes from 16S rRNA gene sequencing using a k-mer based representation of shallow sub-samples
    • Article
    • Peer Reviewed
    UC Berkeley Previously Published Works (2019)
    In this paper, we present peptide-pair encoding (PPE), a general-purpose probabilistic segmentation of protein sequences into commonly occurring variable-length sub-sequences. The idea of PPE segmentation is inspired by the byte-pair encoding (BPE) text compression algorithm, which has recently gained popularity in subword neural machine translation. We modify this algorithm by adding a sampling framework allowing for multiple ways of segmenting a sequence. PPE segmentation steps can be learned over a large set of protein sequences (Swiss-Prot) or even a domain-specific dataset and then applied to a set of unseen sequences. This representation can be widely used as the input to any downstream machine learning tasks in protein bioinformatics. In particular, here, we introduce this representation through protein motif discovery and protein sequence embedding. (i) DiMotif: we present DiMotif as an alignment-free discriminative motif discovery method and evaluate the method for finding protein motifs in three different settings: (1) comparison of DiMotif with two existing approaches on 20 distinct motif discovery problems which are experimentally verified, (2) classification-based approach for the motifs extracted for integrins, integrin-binding proteins, and biofilm formation, and (3) in sequence pattern searching for nuclear localization signal. The DiMotif, in general, obtained high recall scores, while having a comparable F1 score with other methods in the discovery of experimentally verified motifs. Having high recall suggests that the DiMotif can be used for short-list creation for further experimental investigations on motifs. In the classification-based evaluation, the extracted motifs could reliably detect the integrins, integrin-binding, and biofilm formation-related proteins on a reserved set of sequences with high F1 scores. (ii) ProtVecX: we extend k-mer based protein vector (ProtVec) embedding to variablelength protein embedding using PPE sub-sequences. We show that the new method of embedding can marginally outperform ProtVec in enzyme prediction as well as toxin prediction tasks. In addition, we conclude that the embeddings are beneficial in protein classification tasks when they are combined with raw amino acids k-mer features.
      Cover page: Probabilistic variable-length segmentation of protein sequences for discriminative motif discovery (DiMotif) and sequence embedding (ProtVecX)
      • Article
      • Peer Reviewed
      UC Berkeley Previously Published Works (2019)
      The author wishes to inform readers that the affiliations for the authors were incorrect in the original paper. They appear correctly above. The paper has been corrected online.
        Cover page: MicroPheno: predicting environments and host phenotypes from 16S rRNA gene sequencing using a k-mer based representation of shallow sub-samples
        • Article
        • Peer Reviewed
        LBL Publications (2006)

        Metagenome studies have retrieved vast amounts of sequence out of a variety of environments, leading to novel discoveries and great insights into the uncultured microbial world. Except for very simple communities, diversity makes sequence assembly and analysis a very challenging problem. To understand the structure a 5 nd function of microbial communities, a taxonomic characterization of the obtained sequence fragments is highly desirable, yet currently limited mostly to those sequences that contain phylogenetic marker genes. We show that for clades at the rank of domain down to genus, sequence composition allows the very accurate phylogenetic 10 characterization of genomic sequence. We developed a composition-based classifier, PhyloPythia, for de novo phylogenetic sequence characterization and have trained it on adata set of 340 genomes. By extensive evaluation experiments we show that the methodis accurate across all taxonomic ranks considered, even for sequences that originate fromnovel organisms and are as short as 1kb. Application to two metagenome datasets 15 obtained from samples of phosphorus-removing sludge showed that the method allows the accurate classification at genus level of most sequence fragments from the dominant populations, while at the same time correctly characterizing even larger parts of the samples at higher taxonomic levels.

          Cover page: Accurate phylogenetic classification of DNA fragments based on sequence composition
          • Article
          • Peer Reviewed
          Joint Genome Institute (2015)
          Software is now both central and essential to modern biology, yet lack of availability, difficult installations, and complex user interfaces make software hard to obtain and use. Containerisation, as exemplified by the Docker platform, has the potential to solve the problems associated with sharing software. We propose bioboxes: containers with standardised interfaces to make bioinformatics software interchangeable.
            Cover page: Bioboxes: standardised containers for interchangeable bioinformatics software
            • Article
            • Peer Reviewed
            UC San Diego Previously Published Works (2012)
            Here we describe, the longest microbial time-series analyzed to date using high-resolution 16S rRNA tag pyrosequencing of samples taken monthly over 6 years at a temperate marine coastal site off Plymouth, UK. Data treatment effected the estimation of community richness over a 6-year period, whereby 8794 operational taxonomic units (OTUs) were identified using single-linkage preclustering and 21 130 OTUs were identified by denoising the data. The Alphaproteobacteria were the most abundant Class, and the most frequently recorded OTUs were members of the Rickettsiales (SAR 11) and Rhodobacteriales. This near-surface ocean bacterial community showed strong repeatable seasonal patterns, which were defined by winter peaks in diversity across all years. Environmental variables explained far more variation in seasonally predictable bacteria than did data on protists or metazoan biomass. Change in day length alone explains >65% of the variance in community diversity. The results suggested that seasonal changes in environmental variables are more important than trophic interactions. Interestingly, microbial association network analysis showed that correlations in abundance were stronger within bacterial taxa rather than between bacteria and eukaryotes, or between bacteria and environmental variables.
              Cover page: Defining seasonal marine microbial community dynamics.
              • Article
              • Peer Reviewed
              LBL Publications (2016)
              Termite gut flagellates are typically colonized by specific bacterial symbionts. Here we describe the phylogeny, ultrastructure and subcellular location of 'Candidatus Adiutrix intracellularis', an intracellular symbiont of Trichonympha collaris in the termite Zootermopsis nevadensis. It represents a novel, deep-branching clade of uncultured Deltaproteobacteria widely distributed in intestinal tracts of termites and cockroaches. Fluorescence in situ hybridization and transmission electron microscopy localized the endosymbiont near hydrogenosomes in the posterior part and near the ectosymbiont 'Candidatus Desulfovibrio trichonymphae' in the anterior part of the host cell. The draft genome of 'Ca. Adiutrix intracellularis' obtained from a metagenomic library revealed the presence of a complete gene set encoding the Wood-Ljungdahl pathway, including two homologs of fdhF encoding hydrogenase-linked formate dehydrogenases (FDHH ) and all other components of the recently described hydrogen-dependent carbon dioxide reductase (HDCR) complex, which substantiates previous claims that the symbiont is capable of reductive acetogenesis from CO2 and H2 . The close phylogenetic relationship between the HDCR components and their homologs in homoacetogenic Firmicutes and Spirochaetes suggests that the deltaproteobacterium acquired the capacity for homoacetogenesis via lateral gene transfer. The presence of genes for nitrogen fixation and the biosynthesis of amino acids and cofactors indicate the nutritional nature of the symbiosis.
                Cover page: ‘Candidatus Adiutrix intracellularis’, an endosymbiont of termite gut flagellates, is the first representative of a deep‐branching clade of Deltaproteobacteria and a putative homoacetogen
                • Article
                • Peer Reviewed
                UC Berkeley Previously Published Works (2020)
                Limited therapy options due to antibiotic resistance underscore the need for optimization of current diagnostics. In some bacterial species, antimicrobial resistance can be unambiguously predicted based on their genome sequence. In this study, we sequenced the genomes and transcriptomes of 414 drug-resistant clinical Pseudomonas aeruginosa isolates. By training machine learning classifiers on information about the presence or absence of genes, their sequence variation, and expression profiles, we generated predictive models and identified biomarkers of resistance to four commonly administered antimicrobial drugs. Using these data types alone or in combination resulted in high (0.8-0.9) or very high (> 0.9) sensitivity and predictive values. For all drugs except for ciprofloxacin, gene expression information improved diagnostic performance. Our results pave the way for the development of a molecular resistance profiling tool that reliably predicts antimicrobial susceptibility based on genomic and transcriptomic markers. The implementation of a molecular susceptibility test system in routine microbiology diagnostics holds promise to provide earlier and more detailed information on antibiotic resistance profiles of bacterial pathogens and thus could change how physicians treat bacterial infections.
                  Cover page: Predicting antimicrobial resistance in Pseudomonas aeruginosa with machine learning‐enabled molecular diagnostics
                  • Article
                  • Peer Reviewed
                  Lawrence Berkeley National Laboratory (2006)
                    Cover page: Use of simulated data sets to evaluate the fidelity of Metagenomic processing methods
                    • Article
                    • Peer Reviewed
                    LBL Publications (2008)

                    Enhanced Biological Phosphorus Removal (EBPR) is not well understood at the metabolic level despite being one of the best-studied microbially-mediated industrial processes due to its ecological and economic relevance. Here we present a metagenomic analysis of two lab-scale EBPR sludges dominated by the uncultured bacterium, "Candidatus Accumulibacter phosphatis." This analysis resolves several controversies in EBPR metabolic models and provides hypotheses explaining the dominance of A. phosphatis in this habitat, its lifestyle outside EBPR and probable cultivation requirements. Comparison of the same species from different EBPR sludges highlights recent evolutionary dynamics in the A. phosphatis genome that could be linked to mechanisms for environmental adaptation. In spite of an apparent lack of phylogenetic overlap in the flanking communities of the two sludges studied, common functional themes were found, at least one of them complementary to the inferred metabolism of the dominant organism. The present study provides a much-needed blueprint for a systems-level understanding of EBPR and illustrates that metagenomics enables detailed, often novel, insights into even well-studied biological systems.

                      Cover page: Metagenomic analysis of phosphorus removing sludge communities
                      Top