Ultrasensitive sequencing of STR markers utilizing unique molecular identifiers and the SiMSen-Seq method.

Massively parallel sequencing (MPS) is increasingly applied in forensic short tandem repeat (STR) analysis. The presence of stutter artefacts and other PCR or sequencing errors in the MPS-STR data partly limits the detection of low DNA amounts, e.g., in complex mixtures. Unique molecular identifiers (UMIs) have been applied in several scientific fields to reduce noise in sequencing. UMIs consist of a stretch of random nucleotides, a unique barcode for each starting DNA molecule, that is incorporated in the DNA template using either ligation or PCR. The barcode is used to generate consensus reads, thus removing errors. The SiMSen-Seq (Simple, multiplexed, PCR-based barcoding of DNA for sensitive mutation detection using sequencing) method relies on PCR-based introduction of UMIs and includes a sophisticated hairpin design to reduce unspecific primer binding as well as PCR protocol adjustments to further optimize the reaction. In this study, SiMSen-Seq is applied to develop a proof-of-concept seven STR multiplex for MPS library preparation and an associated bioinformatics pipeline. Additionally, machine learning (ML) models were evaluated to further improve UMI allele calling. Overall, the seven STR multiplex resulted in complete detection and concordant alleles for 47 single-source samples at 1 ng input DNA as well as for low-template samples at 62.5 pg input DNA. For twelve challenging mixtures with minor contributions of 10 pg to 150 pg and ratios of 1-15% relative to the major donor, 99.2% of the expected alleles were detected by applying the UMIs in combination with an ML filter. The main impact of UMIs was a substantially lowered number of artefacts as well as reduced stutter ratios, which were generally below 5% of the parental allele. In conclusion, UMI-based STR sequencing opens new means for improved analysis of challenging crime scene samples including complex mixtures.

Treatment Monitoring of a Patient with Synchronous Metastatic Angiosarcoma and Breast Cancer Using ctDNA.

Angiosarcoma is a rare and aggressive type of soft-tissue sarcoma with high propensity to metastasize. For patients with metastatic angiosarcoma, prognosis is dismal and treatment options are limited. To improve the outcomes, identifying patients with poor treatment response at an earlier stage is imperative, enabling alternative therapy. Consequently, there is a need for improved methods and biomarkers for treatment monitoring. Quantification of circulating tumor-DNA (ctDNA) is a promising approach for patient-specific monitoring of treatment response. In this case report, we demonstrate that quantification of ctDNA using SiMSen-Seq was successfully utilized to monitor a patient with metastatic angiosarcoma. By quantifying ctDNA levels using 25 patient-specific mutations in blood plasma throughout surgery and palliative chemotherapy, we predicted the outcome and monitored the clinical response to treatment. This was accomplished despite the additional complexity of the patient having a synchronous breast cancer. The levels of ctDNA showed a superior correlation to the clinical outcome compared with the radiological evaluations. Our data propose a promising approach for personalized biomarker analysis to monitor treatment in angiosarcomas, with potential applicability to other cancers and for patients with synchronous malignancies.

Digital RNA sequencing using unique molecular identifiers enables ultrasensitive RNA mutation analysis.

Mutation analysis is typically performed at the DNA level since most technical approaches are developed for DNA analysis. However, some applications, like transcriptional mutagenesis, RNA editing and gene expression analysis, require RNA analysis. Here, we combine reverse transcription and digital DNA sequencing to enable low error digital RNA sequencing. We evaluate yield, reproducibility, dynamic range and error correction rate for seven different reverse transcription conditions using multiplexed assays. The yield, reproducibility and error rate vary substantially between the specific conditions, where the yield differs 9.9-fold between the best and worst performing condition. Next, we show that error rates similar to DNA sequencing can be achieved for RNA using appropriate reverse transcription conditions, enabling detection of mutant allele frequencies <0.1% at RNA level. We also detect mutations at both DNA and RNA levels in tumor tissue using a breast cancer panel. Finally, we demonstrate that digital RNA sequencing can be applied to liquid biopsies, analyzing cell-free gene transcripts. In conclusion, we demonstrate that digital RNA sequencing is suitable for ultrasensitive RNA mutation analysis, enabling several basic research and clinical applications.

Principles of digital sequencing using unique molecular identifiers.

Massively parallel sequencing technologies have long been used in both basic research and clinical routine. The recent introduction of digital sequencing has made previously challenging applications possible by significantly improving sensitivity and specificity to now allow detection of rare sequence variants, even at single molecule level. Digital sequencing utilizes unique molecular identifiers (UMIs) to minimize sequencing-induced errors and quantification biases. Here, we discuss the principles of UMIs and how they are used in digital sequencing. We outline the properties of different UMI types and the consequences of various UMI approaches in relation to experimental protocols and bioinformatics. Finally, we describe how digital sequencing can be applied in specific research fields, focusing on cancer management where it can be used in screening of asymptomatic individuals, diagnosis, treatment prediction, prognostication, monitoring treatment efficacy and early detection of treatment resistance as well as relapse.

Identification of genomic binding sites and direct target genes for the transcription factor DDIT3/CHOP.

DDIT3 is a tightly regulated basic leucine zipper (bZIP) transcription factor and key regulator in cellular stress responses. It is involved in a variety of pathological conditions and may cause cell cycle block and apoptosis. It is also implicated in differentiation of some specialized cell types and as an oncogene in several types of cancer. DDIT3 was originally believed to act as a dominant-negative inhibitor by forming heterodimers with other bZIP transcription factors, preventing their DNA binding and transactivating functions. DDIT3 has, however, been reported to bind DNA and regulate target genes. Here, we employed ChIP sequencing combined with microarray-based expression analysis to identify direct binding motifs and target genes of DDIT3. The results reveal DDIT3 binding to motifs similar to other bZIP transcription factors, known to form heterodimers with DDIT3. Binding to a class III satellite DNA repeat sequence was also detected. DDIT3 acted as a DNA-binding transcription factor and bound mainly to the promotor region of regulated genes. ChIP sequencing analysis of histone H3K27 methylation and acetylation showed a strong overlap between H3K27-acetylated marks and DDIT3 binding. These results support a role for DDIT3 as a transcriptional regulator of H3K27ac-marked genes in transcriptionally active chromatin.

UMIErrorCorrect and UMIAnalyzer: Software for Consensus Read Generation, Error Correction, and Visualization Using Unique Molecular Identifiers.

BACKGROUND: Targeted sequencing using unique molecular identifiers (UMIs) enables detection of rare variant alleles in challenging applications, such as cell-free DNA analysis from liquid biopsies. Standard bioinformatics pipelines for data processing and variant calling are not adapted for deep-sequencing data containing UMIs, are inflexible, and require multistep workflows or dedicated computing resources. METHODS: We developed a bioinformatics pipeline using Python and an R package for data analysis and visualization. To validate our pipeline, we analyzed cell-free DNA reference material with known mutant allele frequencies (0%, 0.125%, 0.25%, and 1%) and public data sets. RESULTS: We developed UMIErrorCorrect, a bioinformatics pipeline for analyzing sequencing data containing UMIs. UMIErrorCorrect only requires fastq files as inputs and performs alignment, UMI clustering, error correction, and variant calling. We also provide UMIAnalyzer, a graphical user interface, for data mining, visualization, variant interpretation, and report generation. UMIAnalyzer allows the user to adjust analysis parameters and study their effect on variant calling. We demonstrated the flexibility of UMIErrorCorrect by analyzing data from 4 different targeted sequencing protocols. We also show its ability to detect different mutant allele frequencies in standardized cell-free DNA reference material. UMIErrorCorrect outperformed existing pipelines for targeted UMI sequencing data in terms of variant detection sensitivity. CONCLUSIONS: UMIErrorCorrect and UMIAnalyzer are comprehensive and customizable bioinformatics tools that can be applied to any type of library preparation protocol and enrichment chemistry using UMIs. Access to simple, generic, and open-source bioinformatics tools will facilitate the implementation of UMI-based sequencing approaches in basic research and clinical applications.

FET fusion oncoproteins interact with BRD4 and SWI/SNF chromatin remodelling complex subtypes in sarcoma.

FET fusion oncoproteins containing one of the FET (FUS, EWSR1, TAF15) family proteins juxtaposed to alternative transcription-factor partners are characteristic of more than 20 types of sarcoma and leukaemia. FET oncoproteins bind to the SWI/SNF chromatin remodelling complex, which exists in three subtypes: cBAF, PBAF and GBAF/ncBAF. We used comprehensive biochemical analysis to characterize the interactions between FET oncoproteins, SWI/SNF complexes and the transcriptional coactivator BRD4. Here, we report that FET oncoproteins bind all three main SWI/SNF subtypes cBAF, PBAF and GBAF, and that FET oncoproteins interact indirectly with BRD4 via their shared interaction partner SWI/SNF. Furthermore, chromatin immunoprecipitation sequencing and proteomic analysis showed that FET oncoproteins, SWI/SNF components and BRD4 co-localize on chromatin and interact with mediator and RNA Polymerase II. Our results provide a possible molecular mechanism for the FET-fusion-induced oncogenic transcriptional profiles and may lead to novel therapies targeting aberrant SWI/SNF complexes and/or BRD4 in FET-fusion-caused malignancies.

FUS-DDIT3 Fusion Oncoprotein Expression Affects JAK-STAT Signaling in Myxoid Liposarcoma.

Myxoid liposarcoma is one of the most common sarcoma entities characterized by FET fusion oncogenes. Despite a generally favorable prognosis of myxoid liposarcoma, chemotherapy resistance remains a clinical problem. This cancer stem cell property is associated with JAK-STAT signaling, but the link to the myxoid-liposarcoma-specific FET fusion oncogene FUS-DDIT3 is not known. Here, we show that ectopic expression of FUS-DDIT3 resulted in elevated levels of STAT3 and phosphorylated STAT3. RNA sequencing identified 126 genes that were regulated by both FUS-DDIT3 expression and JAK1/2 inhibition using ruxolitinib. Sixty-six of these genes were connected in a protein interaction network. Fifty-three and 29 of these genes were confirmed as FUS-DDIT3 and STAT3 targets, respectively, using public chromatin immunoprecipitation sequencing data sets. Enriched gene sets among the 126 regulated genes included processes related to cytokine signaling, adipocytokine signaling, and chromatin remodeling. We validated CD44 as a target gene of JAK1/2 inhibition and as a potential cancer stem cell marker in myxoid liposarcoma. Finally, we showed that FUS-DDIT3 interacted with phosphorylated STAT3 in association with subunits of the SWI/SNF chromatin remodeling complex and PRC2 repressive complex. Our data show that the function of FUS-DDIT3 is closely connected to JAK-STAT signaling. Detailed deciphering of molecular mechanisms behind tumor progression opens up new avenues for targeted therapies in sarcomas and leukemia characterized by FET fusion oncogenes.

Monitoring Circulating Tumor DNA During Surgical Treatment in Patients with Gastrointestinal Stromal Tumors.

The majority of patients diagnosed with advanced gastrointestinal stromal tumors (GISTs) are successfully treated with a combination of surgery and tyrosine kinase inhibitors (TKIs). However, it remains challenging to monitor treatment efficacy and identify relapse early. Here, we utilized a sequencing strategy based on molecular barcodes and developed a GIST-specific panel to monitor tumor-specific and TKI resistance mutations in cell-free DNA and applied the approach to patients undergoing surgical treatment. Thirty-two patients with GISTs were included, and 161 blood plasma samples were collected and analyzed at routine visits before and after surgery and at the beginning, during, and after surgery. Patients were included regardless of their risk category. Our GIST-specific sequencing approach allowed detection of tumor-specific mutations and TKI resistance mutations with mutant allele frequency < 0.1%. Circulating tumor DNA (ctDNA) was detected in at least one timepoint in nine of 32 patients, ranging from 0.04% to 93% in mutant allele frequency. High-risk patients were more often ctDNA positive than other risk groups (P < 0.05). Patients with detectable ctDNA also displayed higher tumor cell proliferation rates (P < 0.01) and larger tumor sizes (P < 0.01). All patients who were ctDNA positive during surgery became negative after surgery. Finally, in two patients who progressed on TKI treatment, we detected multiple resistance mutations. Our data show that ctDNA may become a clinically useful biomarker in monitoring treatment efficacy in patients with high-risk GISTs and can assist in treatment decision making.

Comprehensive screening of genomic and metagenomic data reveals a large diversity of tetracycline resistance genes.

Tetracyclines are broad-spectrum antibiotics used to prevent or treat a variety of bacterial infections. Resistance is often mediated through mobile resistance genes, which encode one of the three main mechanisms: active efflux, ribosomal target protection or enzymatic degradation. In the last few decades, a large number of new tetracycline-resistance genes have been discovered in clinical settings. These genes are hypothesized to originate from environmental and commensal bacteria, but the diversity of tetracycline-resistance determinants that have not yet been mobilized into pathogens is unknown. In this study, we aimed to characterize the potential tetracycline resistome by screening genomic and metagenomic data for novel resistance genes. By using probabilistic models, we predicted 1254 unique putative tetracycline resistance genes, representing 195 gene families (<70 % amino acid sequence identity), whereof 164 families had not been described previously. Out of 17 predicted genes selected for experimental verification, 7 induced a resistance phenotype in an Escherichia coli host. Several of the predicted genes were located on mobile genetic elements or in regions that indicated mobility, suggesting that they easily can be shared between bacteria. Furthermore, phylogenetic analysis indicated several events of horizontal gene transfer between bacterial phyla. Our results also suggested that acquired efflux pumps originate from proteobacterial species, while ribosomal protection genes have been mobilized from Firmicutes and Actinobacteria. This study significantly expands the knowledge of known and putatively novel tetracycline resistance genes, their mobility and evolutionary history. The study also provides insights into the unknown resistome and genes that may be encountered in clinical settings in the future.

A comprehensive survey of integron-associated genes present in metagenomes.

BACKGROUND: Integrons are genomic elements that mediate horizontal gene transfer by inserting and removing genetic material using site-specific recombination. Integrons are commonly found in bacterial genomes, where they maintain a large and diverse set of genes that plays an important role in adaptation and evolution. Previous studies have started to characterize the wide range of biological functions present in integrons. However, the efforts have so far mainly been limited to genomes from cultivable bacteria and amplicons generated by PCR, thus targeting only a small part of the total integron diversity. Metagenomic data, generated by direct sequencing of environmental and clinical samples, provides a more holistic and unbiased analysis of integron-associated genes. However, the fragmented nature of metagenomic data has previously made such analysis highly challenging. RESULTS: Here, we present a systematic survey of integron-associated genes in metagenomic data. The analysis was based on a newly developed computational method where integron-associated genes were identified by detecting their associated recombination sites. By processing contiguous sequences assembled from more than 10 terabases of metagenomic data, we were able to identify 13,397 unique integron-associated genes. Metagenomes from marine microbial communities had the highest occurrence of integron-associated genes with levels more than 100-fold higher than in the human microbiome. The identified genes had a large functional diversity spanning over several functional classes. Genes associated with defense mechanisms and mobility facilitators were most overrepresented and more than five times as common in integrons compared to other bacterial genes. As many as two thirds of the genes were found to encode proteins of unknown function. Less than 1% of the genes were associated with antibiotic resistance, of which several were novel, previously undescribed, resistance gene variants. CONCLUSIONS: Our results highlight the large functional diversity maintained by integrons present in unculturable bacteria and significantly expands the number of described integron-associated genes.

Sewage from Airplanes Exhibits High Abundance and Diversity of Antibiotic Resistance Genes.

Airplane sanitary facilities are shared by an international audience. We hypothesized the corresponding sewage to be an extraordinary source of antibiotic-resistant bacteria (ARB) and resistance genes (ARG) in terms of diversity and quantity. Accordingly, we analyzed ARG and ARB in airplane-borne sewage using complementary approaches: metagenomics, quantitative polymerase chain reaction (qPCR), and cultivation. For the purpose of comparison, we also quantified ARG and ARB in the inlets of municipal treatment plants with and without connection to airports. As expected, airplane sewage contained an extraordinarily rich set of mobile ARG, and the relative abundances of genes were mostly increased compared to typical raw sewage of municipal origin. Moreover, combined resistance against third-generation cephalosporins, fluorochinolones, and aminoglycosides was unusually common (28.9%) among Escherichia coli isolated from airplane sewage. This percentage exceeds the one reported for German clinical isolates by a factor of 8. Our findings suggest that airplane-borne sewage can effectively contribute to the fast and global spread of antibiotic resistance.

Pharmacology beyond the patient - The environmental risks of human drugs.

BACKGROUND: The presence of pharmaceuticals in the environment is a growing global concern and although environmental risk assessment is required for approval of new drugs in Europe and the USA, the adequacy of the current triggers and the effects-based assessments has been questioned. OBJECTIVE: To provide a comprehensive analysis of all regulatory compliant aquatic ecotoxicity data and evaluate the current triggers and effects-based environmental assessments to facilitate the development of more efficient approaches for pharmaceuticals toxicity testing. METHODS: Publicly-available regulatory compliant ecotoxicity data for drugs targeting human proteins was compiled together with pharmacological information including drug targets, Cmax and lipophilicity. Possible links between these factors and the ecotoxicity data for effects on, growth, mortality and/or reproduction, were evaluated. The environmental risks were then assessed based on a combined analysis of drug toxicity and predicted environmental concentrations based on European patient consumption data. RESULTS: For most (88%) of the of 975 approved small molecule drugs targeting human proteins a complete set of regulatory compliant ecotoxicity data in the public domain was lacking, highlighting the need for both intelligent approaches to prioritize legacy human drugs for a tailored environmental risk assessment and a transparent database that captures environmental data. We show that presence/absence of drug-target orthologues are predictive of susceptible species for the more potent drugs. Drugs that target the endocrine system represent the highest potency and greatest risk. However, for most drugs (>80%) with a full set of ecotoxicity data, risk quotients assuming worst-case exposure assessments were below one in all European countries indicating low environmental risks for the endpoints assessed. CONCLUSION: We believe that the presented analysis can guide improvements to current testing procedures, and provide valuable approaches for prioritising legacy drugs (i.e. those registered before 2006) for further ecotoxicity testing. For drugs where effects of possible concern (e.g. behaviour) are not captured in regulatory tests, additional mechanistic testing may be required to provide the highest confidence for avoiding environmental impacts.

Identification and reconstruction of novel antibiotic resistance genes from metagenomes.

BACKGROUND: Environmental and commensal bacteria maintain a diverse and largely unknown collection of antibiotic resistance genes (ARGs) that, over time, may be mobilized and transferred to pathogens. Metagenomics enables cultivation-independent characterization of bacterial communities but the resulting data is noisy and highly fragmented, severely hampering the identification of previously undescribed ARGs. We have therefore developed fARGene, a method for identification and reconstruction of ARGs directly from shotgun metagenomic data. RESULTS: fARGene uses optimized gene models and can therefore with high accuracy identify previously uncharacterized resistance genes, even if their sequence similarity to known ARGs is low. By performing the analysis directly on the metagenomic fragments, fARGene also circumvents the need for a high-quality assembly. To demonstrate the applicability of fARGene, we reconstructed ß-lactamases from five billion metagenomic reads, resulting in 221 ARGs, of which 58 were previously not reported. Based on 38 ARGs reconstructed by fARGene, experimental verification showed that 81% provided a resistance phenotype in Escherichia coli. Compared to other methods for detecting ARGs in metagenomic data, fARGene has superior sensitivity and the ability to reconstruct previously unknown genes directly from the sequence reads. CONCLUSIONS: We conclude that fARGene provides an efficient and reliable way to explore the unknown resistome in bacterial communities. The method is applicable to any type of ARGs and is freely available via GitHub under the MIT license.

Modelling of zero-inflation improves inference of metagenomic gene count data.

Metagenomics enables the study of gene abundances in complex mixtures of microorganisms and has become a standard methodology for the analysis of the human microbiome. However, gene abundance data is inherently noisy and contains high levels of biological and technical variability as well as an excess of zeros due to non-detected genes. This makes the statistical analysis challenging. In this study, we present a new hierarchical Bayesian model for inference of metagenomic gene abundance data. The model uses a zero-inflated overdispersed Poisson distribution which is able to simultaneously capture the high gene-specific variability as well as zero observations in the data. By analysis of three comprehensive datasets, we show that zero-inflation is common in metagenomic data from the human gut and, if not correctly modelled, it can lead to substantial reductions in statistical power. We also show, by using resampled metagenomic data, that our model has, compared to other methods, a higher and more stable performance for detecting differentially abundant genes. We conclude that proper modelling of the gene-specific variability, including the excess of zeros, is necessary to accurately describe gene abundances in metagenomic data. The proposed model will thus pave the way for new biological insights into the structure of microbial communities.

ECOdrug: a database connecting drugs and conservation of their targets across species.

Pharmaceuticals are designed to interact with specific molecular targets in humans and these targets generally have orthologs in other species. This provides opportunities for the drug discovery community to use alternative model species for drug development. It also means, however, there is potential for mode of action related effects in non-target wildlife species as many pharmaceuticals reach the environment through patient use and manufacturing wastes. Acquiring insight in drug target ortholog predictions across species and taxonomic groups has proven difficult because of the lack of an optimal strategy and because necessary information is spread across multiple and diverse sources and platforms. We introduce a new research platform tool, ECOdrug, that reliably connects drugs to their protein targets across divergent species. It harmonizes ortholog predictions from multiple sources via a simple user interface underpinning critical applications for a wide range of studies in pharmacology, ecotoxicology and comparative evolutionary biology. ECOdrug can be used to identify species with drug targets and identify drugs that interact with those targets. As such, it can be applied to support intelligent targeted drug safety testing by ensuring appropriate and relevant species are selected in ecological risk assessments. ECOdrug is freely accessible and available at: http://www.ecodrug.org.

Identification of 76 novel B1 metallo-ß-lactamases through large-scale screening of genomic and metagenomic data.

BACKGROUND: Metallo-ß-lactamases are bacterial enzymes that provide resistance to carbapenems, the most potent class of antibiotics. These enzymes are commonly encoded on mobile genetic elements, which, together with their broad substrate spectrum and lack of clinically useful inhibitors, make them a particularly problematic class of antibiotic resistance determinants. We hypothesized that there is a large and unexplored reservoir of unknown metallo-ß-lactamases, some of which may spread to pathogens, thereby threatening public health. The aim of this study was to identify novel metallo-ß-lactamases of class B1, the most clinically important subclass of these enzymes. RESULTS: Based on a new computational method using an optimized hidden Markov model, we analyzed over 10,000 bacterial genomes and plasmids together with more than 5 terabases of metagenomic data to identify novel metallo-ß-lactamase genes. In total, 76 novel genes were predicted, forming 59 previously undescribed metallo-ß-lactamase gene families. The ability to hydrolyze imipenem in an Escherichia coli host was experimentally confirmed for 18 of the 21 tested genes. Two of the novel B1 metallo-ß-lactamase genes contained atypical zinc-binding motifs in their active sites, which were previously undescribed for metallo-ß-lactamases. Phylogenetic analysis showed that B1 metallo-ß-lactamases could be divided into five major groups based on their evolutionary origin. Our results also show that, except for one, all of the previously characterized mobile B1 ß-lactamases are likely to have originated from chromosomal genes present in Shewanella spp. and other Proteobacterial species. CONCLUSIONS: This study more than doubles the number of known B1 metallo-ß-lactamases. The findings have further elucidated the diversity and evolutionary history of this important class of antibiotic resistance genes and prepare us for some of the challenges that may be faced in clinics in the future.

Comparison of four DNA extraction methods for comprehensive assessment of 16S rRNA bacterial diversity in marine biofilms using high-throughput sequencing.

High-throughput DNA sequencing technologies are increasingly used for the metagenomic characterisation of microbial biodiversity. However, basic issues, such as the choice of an appropriate DNA extraction method, are still not resolved for non-model microbial communities. This study evaluates four commonly used DNA extraction methods for marine periphyton biofilms in terms of DNA yield, efficiency, purity, integrity and resulting 16S rRNA bacterial diversity. Among the tested methods, the Plant DNAzol® Reagent (PlantDNAzol) and the FastDNA® SPIN Kit for Soil (FastDNA Soil) methods were best suited to extract high quantities of DNA (77-130 µg g wet wt-1). Lower amounts of DNA were obtained (<37 µg g wet wt-1) with the Power Plant® Pro DNA Isolation Kit (PowerPlant) and the Power Biofilm® DNA Isolation Kit (PowerBiofilm) methods, but integrity and purity of the extracted DNA were higher. Results from 16S rRNA amplicon sequencing demonstrate that the choice of a DNA extraction method significantly influences the bacterial community profiles generated. A higher number of bacterial OTUs were detected when DNA was extracted with the PowerBiofilm and the PlantDNAzol methods. Overall, this study demonstrates the potential bias in metagenomic diversity estimates associated with different DNA extraction methods.

HirBin: high-resolution identification of differentially abundant functions in metagenomes.

BACKGROUND: Gene-centric analysis of metagenomics data provides information about the biochemical functions present in a microbiome under a certain condition. The ability to identify significant differences in functions between metagenomes is dependent on accurate classification and quantification of the sequence reads (binning). However, biological effects acting on specific functions may be overlooked if the classes are too general. METHODS: Here we introduce High-Resolution Binning (HirBin), a new method for gene-centric analysis of metagenomes. HirBin combines supervised annotation with unsupervised clustering to bin sequence reads at a higher resolution. The supervised annotation is performed by matching sequence fragments to genes using well-established protein domains, such as TIGRFAM, PFAM or COGs, followed by unsupervised clustering where each functional domain is further divided into sub-bins based on sequence similarity. Finally, differential abundance of the sub-bins is statistically assessed. RESULTS: We show that HirBin is able to identify biological effects that are only present at more specific functional levels. Furthermore we show that changes affecting more specific functional levels are often diluted at the more general level and therefore overlooked when analyzed using standard binning approaches. CONCLUSIONS: HirBin improves the resolution of the gene-centric analysis of metagenomes and facilitates the biological interpretation of the results. HirBin is implemented as a Python package and is freely available for download at http://bioinformatics.math.chalmers.se/hirbin .

Variability in Metagenomic Count Data and Its Influence on the Identification of Differentially Abundant Genes.

Metagenomics is the study of microorganisms in environmental and clinical samples using high-throughput sequencing of random fragments of their DNA. Since metagenomics does not require any prior culturing of isolates, entire microbial communities can be studied directly in their natural state. In metagenomics, the abundance of genes is quantified by sorting and counting the DNA fragments. The resulting count data are high-dimensional and affected by high levels of technical and biological noise that make the statistical analysis challenging. In this article, we introduce an hierarchical overdispersed Poisson model to explore the variability in metagenomic data. By analyzing three comprehensive data sets, we show that the gene-specific variability varies substantially between genes and is dependent on biological function. We also assess the power of identifying differentially abundant genes and show that incorrect assumptions about the gene-specific variability can lead to unacceptable high rates of false positives. Finally, we evaluate shrinkage approaches to improve the variance estimation and show that the prior choice significantly affects the statistical power. The results presented in this study further elucidate the complex variance structure of metagenomic data and provide suggestions for accurate and reliable identification of differentially abundant genes.