riboFrame: An Improved Method for Microbial Taxonomy Profiling from Non-Targeted Metagenomics

Convolutional neural networks improve fungal classification

Sequence classification plays an important role in metagenomics studies. We assess the deep neural network approach for fungal sequence classification as it has emerged as a successful paradigm for big data classification and clustering. Two deep learning-based classifiers, a convolutional neural network (CNN) and a deep belief network (DBN) were trained using our recently released barcode datasets. Experimental results show that CNN outperformed the traditional BLAST classification and the most accurate machine learning based Ribosomal Database Project (RDP) classifier on datasets that had many of the labels present in the training datasets. When classifying an independent dataset namely the "Top 50 Most Wanted Fungi", CNN and DBN assigned less sequences than BLAST. However, they could assign much more sequences than the RDP classifier. In terms of efficiency, it took the machine learning classifiers up to two seconds to classify a test dataset while it was 53 s for BLAST. The result of the current study will enable us to speed up the taxonomic assignments for the fungal barcode sequences generated at our institute as ~ 70% of them still need to be validated for public release. In addition, it will help to quickly provide a taxonomic profile for metagenomics samples.

DAIRYdb: a manually curated reference database for improved taxonomy annotation of 16S rRNA gene sequences from dairy products

Background

Reads assignment to taxonomic units is a key step in microbiome analysis pipelines. To date, accurate taxonomy annotation of 16S reads, particularly at species rank, is still challenging due to the short size of read sequences and differently curated classification databases. The close phylogenetic relationship between species encountered in dairy products, however, makes it crucial to annotate species accurately to achieve sufficient phylogenetic resolution for further downstream ecological studies or for food diagnostics. Curated databases dedicated to the environment of interest are expected to improve the accuracy and resolution of taxonomy annotation.

Results

We provide a manually curated database composed of 10’290 full-length 16S rRNA gene sequences from prokaryotes tailored for dairy products analysis (https://github.com/marcomeola/DAIRYdb).

The performance of the DAIRYdb was compared with the universal databases Silva, LTP, RDP and Greengenes. The DAIRYdb significantly outperformed all other databases independently of the classification algorithm by enabling higher accurate taxonomy annotation down to the species rank. The DAIRYdb accurately annotates over 90% of the sequences of either single or paired hypervariable regions automatically.

The manually curated DAIRYdb strongly improves taxonomic annotation accuracy for microbiome studies in dairy environments. The DAIRYdb is a practical solution that enables automatization of this key step, thus facilitating the routine application of NGS microbiome analyses for microbial ecology studies and diagnostics in dairy products.

Electronic supplementary material

The online version of this article (10.1186/s12864-019-5914-8) contains supplementary material, which is available to authorized users.

Metagenomic Composition Analysis of an Ancient Sequenced Polar Bear Jawbone from Svalbard

The sequencing of ancient DNA samples provides a novel way to find, characterize, and distinguish exogenous genomes of endogenous targets. After sequencing, computational composition analysis enables filtering of undesired sources in the focal organism, with the purpose of improving the quality of assemblies and subsequent data analysis. More importantly, such analysis allows extinct and extant species to be identified without requiring a specific or new sequencing run. However, the identification of exogenous organisms is a complex task, given the nature and degradation of the samples, and the evident necessity of using efficient computational tools, which rely on algorithms that are both fast and highly sensitive. In this work, we relied on a fast and highly sensitive tool, FALCON-meta, which measures similarity against whole-genome reference databases, to analyse the metagenomic composition of an ancient polar bear (Ursus maritimus) jawbone fossil. The fossil was collected in Svalbard, Norway, and has an estimated age of 110,000 to 130,000 years. The FASTQ samples contained 349 GB of nonamplified shotgun sequencing data. We identified and localized, relative to the FASTQ samples, the genomes with significant similarities to reference microbial genomes, including those of viruses, bacteria, and archaea, and to fungal, mitochondrial, and plastidial sequences. Among other striking features, we found significant similarities between modern-human, some bacterial and viral sequences (contamination) and the organelle sequences of wild carrot and tomato relative to the whole samples. For each exogenous candidate, we ran a damage pattern analysis, which in addition to revealing shallow levels of damage in the plant candidates, identified the source as contamination.

Deep learning models for bacteria taxonomic classification of metagenomic data

Background

An open challenge in translational bioinformatics is the analysis of sequenced metagenomes from various environmental samples. Of course, several studies demonstrated the 16S ribosomal RNA could be considered as a barcode for bacteria classification at the genus level, but till now it is hard to identify the correct composition of metagenomic data from RNA-seq short-read data. 16S short-read data are generated using two next generation sequencing technologies, i.e. whole genome shotgun (WGS) and amplicon (AMP); typically, the former is filtered to obtain short-reads belonging to a 16S shotgun (SG), whereas the latter take into account only some specific 16S hypervariable regions. The above mentioned two sequencing technologies, SG and AMP, are used alternatively, for this reason in this work we propose a deep learning approach for taxonomic classification of metagenomic data, that can be employed for both of them.

Results

To test the proposed pipeline, we simulated both SG and AMP short-reads, from 1000 16S full-length sequences. Then, we adopted a k-mer representation to map sequences as vectors into a numerical space. Finally, we trained two different deep learning architecture, i.e., convolutional neural network (CNN) and deep belief network (DBN), obtaining a trained model for each taxon. We tested our proposed methodology to find the best parameters configuration, and we compared our results against the classification performances provided by a reference classifier for bacteria identification, known as RDP classifier. We outperformed the RDP classifier at each taxonomic level with both architectures. For instance, at the genus level, both CNN and DBN reached 91.3% of accuracy with AMP short-reads, whereas RDP classifier obtained 83.8% with the same data.

Conclusions

In this work, we proposed a 16S short-read sequences classification technique based on k-mer representation and deep learning architecture, in which each taxon (from phylum to genus) generates a classification model. Experimental results confirm the proposed pipeline as a valid approach for classifying bacteria sequences; for this reason, our approach could be integrated into the most common tools for metagenomic analysis. According to obtained results, it can be successfully used for classifying both SG and AMP data.

Electronic supplementary material

The online version of this article (10.1186/s12859-018-2182-6) contains supplementary material, which is available to authorized users.

Metagenomic Approaches to Investigate the Contribution of the Vineyard Environment to the Quality of Wine Fermentation: Potentials and Difficulties

The winemaking is a complex process that begins in the vineyard and ends at consumption moment. Recent reports have shown the relevance of microbial populations in the definition of the regional organoleptic and sensory characteristics of a wine. Metagenomic approaches, allowing the exhaustive identification of microorganisms present in complex samples, have recently played a fundamental role in the dissection of the contribution of the vineyard environment to wine fermentation. Systematic approaches have explored the impact of agronomical techniques, vineyard topologies, and climatic changes on bacterial and fungal populations found in the vineyard and in fermentations, also trying to predict or extrapolate the effects on the sensorial characteristics of the resulting wine. This review is aimed at highlighting the major technical and experimental challenges in dissecting the contribution of the vineyard and native environments microbiota to the wine fermentation process, and how metagenomic approaches can help in understanding microbial fluxes and selections across the environments and specimens related to wine fermentation.

Environmental drivers of a microbial genomic transition zone in the ocean's interior

The core properties of microbial genomes, including GC content and genome size, are known to vary widely among different bacteria and archaea ^1,2 . Several hypotheses have been proposed to explain this genomic variability, but the fundamental drivers that shape bacterial and archaeal genomic properties remain uncertain ^3-7 . Here, we report the existence of a sharp genomic transition zone below the photic zone, where bacterial and archaeal genomes and proteomes undergo a community-wide punctuated shift. Across a narrow range of increasing depth of just tens of metres, diverse microbial clades trend towards larger genome size, higher genomic GC content, and proteins with higher nitrogen but lower carbon content. These community-wide changes in genome features appear to be driven by gradients in the surrounding environmental energy and nutrient fields. Collectively, our data support hypotheses invoking nutrient limitation as a central driver in the evolution of core bacterial and archaeal genomic and proteomic properties.

Single-virus genomics reveals hidden cosmopolitan and abundant viruses

Microbes drive ecosystems under constraints imposed by viruses. However, a lack of virus genome information hinders our ability to answer fundamental, biological questions concerning microbial communities. Here we apply single-virus genomics (SVGs) to assess whether portions of marine viral communities are missed by current techniques. The majority of the here-identified 44 viral single-amplified genomes (vSAGs) are more abundant in global ocean virome data sets than published metagenome-assembled viral genomes or isolates. This indicates that vSAGs likely best represent the dsDNA viral populations dominating the oceans. Species-specific recruitment patterns and virome simulation data suggest that vSAGs are highly microdiverse and that microdiversity hinders the metagenomic assembly, which could explain why their genomes have not been identified before. Altogether, SVGs enable the discovery of some of the likely most abundant and ecologically relevant marine viral species, such as vSAG 37-F6, which were overlooked by other methodologies.

RiboTagger: fast and unbiased 16S/18S profiling using whole community shotgun metagenomic or metatranscriptome surveys

Background

Taxonomic profiling of microbial communities is often performed using small subunit ribosomal RNA (SSU) amplicon sequencing (16S or 18S), while environmental shotgun sequencing is often focused on functional analysis. Large shotgun datasets contain a significant number of SSU sequences and these can be exploited to perform an unbiased SSU--based taxonomic analysis.

Results

Here we present a new program called RiboTagger that identifies and extracts taxonomically informative ribotags located in a specified variable region of the SSU gene in a high-throughput fashion.

Conclusions

RiboTagger permits fast recovery of SSU-RNA sequences from shotgun nucleic acid surveys of complex microbial communities. The program targets all three domains of life, exhibits high sensitivity and specificity and is substantially faster than comparable programs.