Mock Data Sets Research Articles

Power spectrum modelling of galaxy and radio intensity maps including observational effects

ABSTRACT Fluctuations in the large-scale structure of the Universe contain significant information about cosmological physics, but are modulated in survey data sets by various observational effects. Building on existing literature, we provide a general treatment of how fluctuation power spectra are modified by a position-dependent selection function, noise, weighting, smoothing, pixelization, and discretization. Our work has relevance for the spatial power spectrum analysis of galaxy surveys with spectroscopic or accurate photometric redshifts, and radio intensity-mapping surveys of the sky brightness temperature including generic noise, telescope beams, and pixelization. We consider the autopower spectrum of a field, the cross-power spectrum between two fields and the multipoles of these power spectra with respect to a curved sky, deriving the corresponding power spectrum models, estimators, errors, and optimal weights. We note that ‘FKP weights’ for individual tracers do not in general provide the optimal weights when measuring the cross-power spectrum. We validate our models using mock data sets drawn fromN-body simulations.1 Our treatment should be useful for modelling and studying cosmological fluctuation fields in observed and simulated data sets.

Do reverberation mapping analyses provide an accurate picture of the broad-line region?

Abstract Reverberation mapping (RM) is a powerful approach for determining the nature of the broad-line region (BLR) in active galactic nuclei. However, inferring physical BLR properties from an observed spectroscopic time series is a difficult inverse problem. Here, we present a blind test of two widely used RM methods: MEMEcho (developed by Horne) and CARAMEL (developed by Pancoast and collaborators). The test data are simulated spectroscopic time series that track the Hα emission line response to an empirical continuum light curve. The underlying BLR model is a rotating, biconical accretion disc wind, and the synthetic spectra are generated via self-consistent ionization and radiative transfer simulations. We generate two mock data sets, representing Seyfert galaxies and QSOs. The Seyfert model produces a largely negative response, which neither method can recover. However, both fail ‘gracefully', neither generating spurious results. For the QSO model both CARAMEL and expert interpretation of MEMEchoś output both capture the broadly annular, rotation-dominated nature of the line-forming region, though MEMEcho analysis overestimates its size by 50 per cent, but CARAMEL is unable to distinguish between additional inflow and outflow components. Despite fitting individual spectra well, the CARAMEL velocity-delay maps and RMS line profiles are strongly inconsistent with the input data. Finally, since the Hα line-forming region is rotation dominated, neither method recovers the disc wind nature of the underlying BLR model. Thus considerable care is required when interpreting the results of RM analyses in terms of physical models.

Schwarzschild modeling of barred galaxies

AbstractWe review the Schwarzschild orbit-superposition approach and present a new implementation of this method, which can deal with a large class of systems, including rotating barred disk galaxies. We discuss two conceptual problems in this field: the intrinsic degeneracy of determining the potential from line-of-sight kinematics, and the non-uniqueness of deprojection and related biases in potential inference, especially acute for triaxial bars. When applied to mock datasets with known 3d shape, our method correctly recovers the pattern speed and other potential parameters. However, more work is needed to systematically address these two problems for real observational datasets.

Evaluating the ability of triaxial Schwarzschild modelling to estimate properties of galaxies from the Illustris simulation

We evaluate the capabilities of Schwarzschild's orbit-superposition method by applying it to galaxies from the large scale, high resolution Illustris simulation. Nine early-type galaxies with a range of triaxiality are selected, and we create mock integral field unit data for five line-of-sight projections of each galaxy. Each of the 45 mock data sets is taken as an independent observed galaxy. Using van den Bosch's 2008 triaxial Schwarzschild implementation, we assess model estimates of various galaxy properties, covering mass profiles, intrinsic shapes, stellar orbit distributions and velocity anisotropies. Total mass within $\overline{R_{\rm e}}$ is recovered well with average deviations within $\pm15$ percent. Stellar mass is underestimated by $\sim24$ percent and dark matter overestimated by $\sim38$ percent (assuming an NFW dark matter profile and allowing for degeneracy between stellar mass and dark matter mass). Using a gNFW profile, these values improve to $\sim13$ percent for stellar mass and $\sim18$ percent for dark matter. Axis ratio estimates show a moderate bias of $\Delta (b/a)=0.07$ and $\Delta (c/a)=0.14$ ($a\ge b\ge c$). Distributions of the orbit circularities $\lambda_z$ and $\lambda_x$, representing rotation about the minor and major axes, are well reconstructed. Separating orbits into thermal categories, our models match the average fractions of these categories to within $10$ percent. Velocity anisotropy is well estimated with values matching in the inner regions but becoming slightly radially biased in the outer regions. Overall, the galaxy property estimates we obtained using Schwarzschild modelling are not implausible and are representative of the simulated galaxies we modelled.

Explicit Bayesian treatment of unknown foreground contaminations in galaxy surveys

The treatment of unknown foreground contaminations will be one of the major challenges for galaxy clustering analyses of coming decadal surveys. These data contaminations introduce erroneous large-scale effects in recovered power spectra and inferred dark matter density fields. In this work, we present an effective solution to this problem in the form of a robust likelihood designed to account for effects due to unknown foreground and target contaminations. Conceptually, this robust likelihood marginalizes over the unknown large-scale contamination amplitudes. We showcase the effectiveness of this novel likelihood via an application to a mock SDSS-III data set subject to dust extinction contamination. In order to illustrate the performance of our proposed likelihood, we infer the underlying dark-matter density field and reconstruct the matter power spectrum, being maximally agnostic about the foregrounds. The results are compared to those of an analysis with a standard Poissonian likelihood, as typically used in modern large-scale structure analyses. While the standard Poissonian analysis yields excessive power for large-scale modes and introduces an overall bias in the power spectrum, our likelihood provides unbiased estimates of the matter power spectrum over the entire range of Fourier modes considered in this work. Further, we demonstrate that our approach accurately accounts for and corrects the effects of unknown foreground contaminations when inferring three-dimensional density fields. Robust likelihood approaches, as presented in this work, will be crucial to control unknown systematic error and maximize the outcome of the decadal surveys.

Mixing-length calibration from field stars

Aims. We critically analysed the theoretical foundation and statistical reliability of the mixing-length calibration by means of standard (Teff, [Fe/H]) and global asteroseismic observables (Δν,νmax) of field stars. We also discussed the soundness of inferring a possible metallicity dependence of the mixing-length parameter from field stars.Methods. We followed a theoretical approach based on mock datasets of artificial stars sampled from a grid of stellar models with a fixed mixing-length parameterαml. We then recovered the mixing-length parameter of the mock stars by means of SCEPtER maximum-likelihood algorithm. We finally analysed the differences between the true and recovered mixing-length values quantifying the random errors due to the observational uncertainties and the biases due to possible discrepancies in the chemical composition and input physics between artificial stars and the models adopted in the recovery.Results. We verified that theαmlestimates are affected by a huge spread, even in the ideal configuration of perfect agreement between the mock data and the recovery grid of models. While the artificial stars were computed at fixed solar-calibratedαml = 2.10, the recovered values had a mean of 2.20 and a standard deviation of 0.52. Then we explored the case in which the solar heavy-element mixture used to compute the models is different from that adopted in the artificial stars. We found an estimated mixing-length mean of 2.24 ± 0.48 and, more interestingly, a metallicity relationship in whichαmlincreases by 0.4 for an increase of 1 dex in [Fe/H]. Thus, a simple heavy-element mixture mismatch induced a spurious, but statistically robust, dependence of the estimated mixing-length on metallicity. The origin of this trend was further investigated considering the differences in the initial helium abundanceY– [Fe/H] – initial metallicityZrelation assumed in the models and data. We found that a discrepancy between the adopted helium-to-metal enrichment ratio ΔY/ΔZcaused the appearance of spurious trends in the estimated mixing-length values. An underestimation of its value from ΔY/ΔZ = 2.0 in the mock data to ΔY/ΔZ = 1.0 in the recovery grid resulted in an increasing trend, while the opposite behaviour occurred for an equivalent overestimation. A similar effect was caused by an offset in the [Fe/H] to global metallicityZconversion. A systematic overestimation of [Fe/H] by 0.1 dex in the recovery grid of models forced an increasing trend ofαmlversus [Fe/H] of about 0.2 per dex. We also explored the impact of some possible discrepancies between the adopted input physics in the recovery grid of models and mock data. We observed an induced trend with metallicity of about Δαml = 0.3 per dex when the effect of the microscopic diffusion is neglected in the recovery grid, while no trends originated from a wrong assumption on the effective temperature scale by ±100 K. Finally, we proved that the impact of different assumptions on the outer boundary conditions was apparent only in the RGB phase.Conclusions. We showed that the mixing-length estimates of field stars are affected by a huge spread even in an ideal case in which the stellar models used to estimateαmlare exactly the same models as used to build the mock dataset. Moreover, we proved that there are many assumptions adopted in the stellar models used in the calibration that can induce spurious trend of the estimatedαmlwith [Fe/H]. Therefore, any attempt to calibrate the mixing-length parameter by means ofTeff, [Fe/H], Δν, andνmaxof field stars seems to be statistically poorly reliable. As such, any claim about the possible dependence of the mixing-length on the metallicity for field stars should be considered cautiously and critically.

On the degeneracy between baryon feedback and massive neutrinos as probed by matter clustering and weak lensing

Massive neutrinos, due to their free streaming, produce a suppression in the matter power spectrum at intermediate and small scales which could be probed by galaxy clustering and/or weak lensing observables. This effect happens at scales that are also influenced by baryon feedback, i.e. galactic winds or Active Galactic Nuclei (AGN) feedback, which in realistic hydrodynamic simulations has also been shown to produce a suppression of power. Leaving aside, for the moment, the complex issue of galaxy bias, we focus here on matter clustering and tomographic weak lensing, we investigate the possible degeneracy between baryon feedback and neutrinos showing that it is not likely to degrade significantly the measurement of neutrino mass in future surveys. To do so, we generate mock data sets and fit them using the Markov Chain Monte Carlo (MCMC) technique and explore degeneracies between feedback parameters and neutrino mass. We model baryon feedback through fitting functions, while massive neutrinos are accounted for, also in the non-linear regime, using HALOFIT calibrated against accurate N-body neutrino simulations. In the error budget, we include the uncertainty in the modelling of non-linearities. For both matter clustering and weak lensing, we always recover the input neutrino mass within ∼0.25σ confidence level. Finally, we also take into account the intrinsic alignment effect in the weak lensing mock data. Even in this case, we are able to recover the right parameters: in particular, we find a significant degeneracy pattern between Mν and the intrinsic alignment parameter AIA.

Denoising, deconvolving, and decomposing multi-domain photon observations

Astronomical imaging based on photon count data is a non-trivial task. In this context we show how to denoise, deconvolve, and decompose multi-domain photon observations. The primary objective is to incorporate accurate and well motivated likelihood and prior models in order to give reliable estimates about morphologically different but superimposed photon flux components present in the data set. Thereby we denoise and deconvolve photon counts, while simultaneously decomposing them into diffuse, point-like and uninteresting background radiation fluxes. The decomposition is based on a probabilistic hierarchical Bayesian parameter model within the framework of information field theory (IFT). In contrast to its predecessor D3PO, D4PO reconstructs multi-domain components. Thereby each component is defined over its own direct product of multiple independent domains, for example location and energy. D4PO has the capability to reconstruct correlation structures over each of the sub-domains of a component separately. Thereby the inferred correlations implicitly define the morphologically different source components, except for the spatial correlations of the point-like flux. Point-like source fluxes are spatially uncorrelated by definition. The capabilities of the algorithm are demonstrated by means of a synthetic, but realistic, mock data set, providing spectral and spatial information about each detected photon. D4PO successfully denoised, deconvolved, and decomposed a photon count image into diffuse, point-like and background flux, each being functions of location as well as energy. Moreover, uncertainty estimates of the reconstructed fields as well as of their correlation structure are provided employing their posterior density function and accounting for the manifolds the domains reside on.

Internal robustness of growth rate data

We perform an Internal Robustness analysis (iR) to a compilation of the most recent $f\sigma_8(z)$ data, using the framework of 1209.1897. The method analyzes combinations of subsets in the data set in a Bayesian model comparison way, potentially finding outliers, subsets of data affected by systematics or new physics. In order to validate our analysis and assess its sensitivity we performed several cross-checks, for example by removing some of the data or by adding artificially contaminated points, while we also generated mock data sets in order to estimate confidence regions of the iR. Applying this methodology, we found no anomalous behavior in the $f\sigma_8(z)$ data set, thus validating its internal robustness.

Searching for a possible dipole anisotropy in acceleration scale with 147 rotationally supported galaxies**Supported by National Natural Science Foundation of China (11675182, 11690022, 11603005)

We report a possible dipole anisotropy in acceleration scale g† with 147 rotationally supported galaxies in the local Universe. It is found that a monopole and dipole correction for the radial acceleration relation can better describe the SPARC data set. The monopole term is negligible but the dipole magnitude is significant. It is also found that the dipole correction is mostly induced by anisotropy in the acceleration scale. The magnitude of the -dipole reaches 0.25 ± 0.04, and its direction is aligned to (l,b) = (171.30° ± 7.18°, −15.41° ± 4.87°), which is very close to the maximum anisotropy direction from the hemisphere comparison method. Furthermore, a robust check shows that the dipole anisotropy could not be reproduced by an isotropic mock data set. However, it is still premature to claim that the Universe is anisotropic, due to the small data samples and uncertainty in the current observations.

Image Reconstruction Techniques in Neutron and Gamma Ray Spectroscopy: Improving Lunar Prospector Data

AbstractWe present improved resolution maps of the Lunar Prospector Neutron Spectrometer thermal, epithermal, and fast neutron data and gamma ray spectrometer Th‐line fluxes via global application of pixon image reconstruction techniques. With the use of mock data sets, we show that the pixon image reconstruction method compares favorably with other methods that have been used in planetary neutron and gamma ray spectroscopy. The improved thermal neutron maps are able to clearly distinguish variations in composition across the lunar surface, including within the lunar basins of Hertzsprung and Schrödinger. The improvement in resolution reveals a correlation between albedo and thermal neutron flux within the basins. The consequent increase in dynamic range confirms that Hertzsprung basin contains one of the most anorthositic parts of the lunar crust, including nearly pure anorthite over a region tens of kilometers in diameter. At Orientale, the reconstructed epithermal neutron data show broad overlap with circular polarization ratio but there remains a mismatch between measures of regolith maturity that sample the surface and those that probe the near subsurface, which is consistent with a complex layering scenario.

Approximating Photo-z PDFs for Large Surveys

Abstract Modern galaxy surveys produce redshift probability density functions (PDFs) in addition to traditional photometric redshift (photo-z) point estimates. However, the storage of photo-z PDFs may present a challenge with increasingly large catalogs, as we face a trade-off between the accuracy of subsequent science measurements and the limitation of finite storage resources. This paper presents qp, a Python package for manipulating parameterizations of one-dimensional PDFs, as suitable for photo-z PDF compression. We use qp to investigate the performance of three simple PDF storage formats (quantiles, samples, and step functions) as a function of the number of stored parameters on two realistic mock data sets, representative of upcoming surveys with different data qualities. We propose some best practices for choosing a photo-z PDF approximation scheme and demonstrate the approach on a science case using performance metrics on both ensembles of individual photo-z PDFs and an estimator of the overall redshift distribution function. We show that both the properties of the set of PDFs we wish to approximate and the fidelity metric(s) chosen affect the optimal parameterization. Additionally, we find that quantiles and samples outperform step functions, and we encourage further consideration of these formats for PDF approximation.

Super-Eddington accreting massive black holes explore high- z cosmology: Monte-Carlo simulations

In this paper, we simulate Super-Eddington accreting massive black holes (SEAMBHs) as the candles to probe cosmology for the first time. SEAMBHs have been demonstrated to be able to provide a new tool for estimating cosmological distance. Thus, we create a series of mock data sets of SEAMBHs, especially in the high redshift region, to check their abilities to probe the cosmology. To fulfill the potential of the SEAMBHs on the cosmology, we apply the simulated data to three projects. The first is the exploration of their abilities to constrain the cosmological parameters, in which we combine different data sets of current observations such as the cosmic microwave background from {\it Planck} and type Ia supernovae from Joint Light-curve Analysis (JLA). We find that the high redshift SEAMBHs can help to break the degeneracies of the background cosmological parameters constrained by {\it Planck} and JLA, thus giving much tighter constraints of the cosmological parameters. The second uses the high redshift SEAMBHs as the complements of the low redshift JLA to constrain the early expansion rate and the dark energy density evolution in the cold dark matter frame. Our results show that these high redshift SEAMBHs are very powerful on constraining the early Hubble rate and the evolution of the dark energy density; thus they can give us more information about the expansion history of our Universe, which is also crucial for testing the $\Lambda$CDM model in the high redshift region. Finally, we check the SEAMBH candles' abilities to reconstruct the equation of state of dark energy at high redshift. In summary, our results show that the SEAMBHs, as the rare candles in the high redshift region, can provide us a new and independent observation to probe cosmology in the future.

Informative graphing of continuous safety variables relative to normal reference limits

BackgroundInterpreting graphs of continuous safety variables can be complicated because differences in age, gender, and testing site methodologies data may give rise to multiple reference limits. Furthermore, data below the lower limit of normal are compressed relative to those points above the upper limit of normal. The objective of this study is to develop a graphing technique that addresses these issues and is visually intuitive.MethodsA mock dataset with multiple reference ranges is initially used to develop the graphing technique. Formulas are developed for conditions where data are above the upper limit of normal, normal, below the lower limit of normal, and below the lower limit of normal when the data value equals zero. After the formulae are developed, an anonymized dataset from an actual set of trials for an approved drug is evaluated comparing the technique developed in this study to standard graphical methods.ResultsFormulas are derived for the novel graphing method based on multiples of the normal limits. The formula for values scaled between the upper and lower limits of normal is a novel application of a readily available scaling formula. The formula for the lower limit of normal is novel and addresses the issue of this value potentially being indeterminate when the result to be scaled as a multiple is zero.ConclusionsThe formulae and graphing method described in this study provides a visually intuitive method to graph continuous safety data including laboratory values, vital sign data.

Modelling non-linear effects of dark energy

We investigate the capabilities of perturbation theory in capturing non-linear effects of dark energy. We test constant and evolving w models, as well as models involving momentum exchange between dark energy and dark matter. Specifically, we compare perturbative predictions at 1-loop level against N-body results for four non-standard equations of state as well as varying degrees of momentum exchange between dark energy and dark matter. The interaction is modelled phenomenologically using a time dependent drag term in the Euler equation. We make comparisons at the level of the matter power spectrum and the redshift space monopole and quadrupole. The multipoles are modelled using the Taruya, Nishimichi and Saito (TNS) redshift space spectrum. We find perturbation theory does very well in capturing non-linear effects coming from dark sector interaction. We isolate and quantify the 1-loop contribution coming from the interaction and from the non-standard equation of state. We find the interaction parameter ξ amplifies scale dependent signatures in the range of scales considered. Non-standard equations of state also give scale dependent signatures within this same regime. In redshift space the match with N-body is improved at smaller scales by the addition of the TNS free parameter σv. To quantify the importance of modelling the interaction, we create mock data sets for varying values of ξ using perturbation theory. This data is given errors typical of Stage IV surveys. We then perform a likelihood analysis using the first two multipoles on these sets and a ξ=0 modelling, ignoring the interaction. We find the fiducial growth parameter f is generally recovered even for very large values of ξ both at z=0.5 and z=1. The ξ=0 modelling is most biased in its estimation of f for the phantom w=−1.1 case.

Planck's view on the spectrum of the Sunyaev–Zeldovich effect

We present a detailed analysis of the stacked frequency spectrum of a large sample of galaxy clusters using Planck data, together with auxiliary data from the AKARI and IRAS missions. Our primary goal is to search for the imprint of relativistic corrections to the thermal Sunyaev-Zeldovich effect (tSZ) spectrum, which allow to measure the temperature of the intracluster medium. We remove Galactic and extragalactic foregrounds with a matched filtering technique, which is validated using simulations with realistic mock data sets. The extracted spectra show the tSZ signal at high significance and reveal an additional far-infrared (FIR) excess, which we attribute to thermal emission from the galaxy clusters themselves. This excess FIR emission from clusters is accounted for in our spectral model. We are able to measure the tSZ relativistic corrections at $2.2\sigma$ by constraining the mean temperature of our cluster sample to $4.4^{+2.1}_{-2.0} \, \mathrm{keV}$. We repeat the same analysis on a subsample containing only the 100 hottest clusters, for which we measure the mean temperature to be $6.0^{+3.8}_{-2.9} \, \mathrm{keV}$, corresponding to $2.0\sigma$. The temperature of the emitting dust grains in our FIR model is constrained to $\simeq 20 \, \mathrm{K}$, consistent with previous studies. Control for systematic biases is done by fitting mock clusters, from which we also show that using the non-relativistic spectrum for SZ signal extraction will lead to a bias in the integrated Compton parameter $Y$, which can be up to 14% for the most massive clusters. We conclude by providing an outlook for the upcoming CCAT-prime telescope, which will improve upon Planck with lower noise and better spatial resolution.

The globular cluster systems of 54 Coma ultra-diffuse galaxies: statistical constraints from HST data

We use data from the HST Coma Cluster Treasury program to assess the richness of the Globular Cluster Systems (GCSs) of 54 Coma ultra-diffuse galaxies (UDGs), 18 of which have a half-light radius exceeding 1.5 kpc. We use a hierarchical Bayesian method tested on a large number of mock datasets to account consistently for the high and spatially varying background counts in Coma. These include both background galaxies and intra-cluster GCs (ICGCs), which are disentangled from the population of member GCs in a probabilistic fashion. We find no candidate for a GCS as rich as that of the Milky Way, our sample has GCSs typical of dwarf galaxies. For the standard relation between GCS richness and halo mass, 33 galaxies have a virial mass $M_{vir}\leq10^{11}M_\odot$ at 90% probability. Only three have $M_{vir}>10^{11}M_\odot$ with the same confidence. The mean colour and spread in colour of the UDG GCs are indistinguishable from those of the abundant population of ICGCs. The majority of UDGs in our sample are consistent with the relation between stellar mass and GC richness of 'normal' dwarf galaxies. Nine systems, however, display GCSs that are richer by a factor of 3 or more (at 90% probability). Six of these have sizes $\lesssim1.4$ kpc. Our results imply that the physical mechanisms responsible for the extended size of the UDGs and for the enhanced GC richness of some cluster dwarfs are at most weakly correlated.

Forecast and analysis of the cosmological redshift drift

The cosmological redshift drift could lead to the next step in high-precision cosmic geometric observations, becoming a direct and irrefutable test for cosmic acceleration. In order to test the viability and possible properties of this effect, also called Sandage–Loeb (SL) test, we generate a model-independent mock data set in order to compare its constraining power with that of the future mock data sets of Type Ia Supernovae (SNe) and Baryon Acoustic Oscillations (BAO). The performance of those data sets is analyzed by testing several cosmological models with the Markov chain Monte Carlo (MCMC) method, both independently as well as combining all data sets. Final results show that, in general, SL data sets allow for remarkable constraints on the matter density parameter today varOmega _m on every tested model, showing also a great complementarity with SNe and BAO data regarding dark energy parameters.

MEGAN-LR: new algorithms allow accurate binning and easy interactive exploration of metagenomic long reads and contigs

BackgroundThere are numerous computational tools for taxonomic or functional analysis of microbiome samples, optimized to run on hundreds of millions of short, high quality sequencing reads. Programs such as MEGAN allow the user to interactively navigate these large datasets. Long read sequencing technologies continue to improve and produce increasing numbers of longer reads (of varying lengths in the range of 10k-1M bps, say), but of low quality. There is an increasing interest in using long reads in microbiome sequencing, and there is a need to adapt short read tools to long read datasets.MethodsWe describe a new LCA-based algorithm for taxonomic binning, and an interval-tree based algorithm for functional binning, that are explicitly designed for long reads and assembled contigs. We provide a new interactive tool for investigating the alignment of long reads against reference sequences. For taxonomic and functional binning, we propose to use LAST to compare long reads against the NCBI-nr protein reference database so as to obtain frame-shift aware alignments, and then to process the results using our new methods.ResultsAll presented methods are implemented in the open source edition of MEGAN, and we refer to this new extension as MEGAN-LR (MEGAN long read). We evaluate the LAST+MEGAN-LR approach in a simulation study, and on a number of mock community datasets consisting of Nanopore reads, PacBio reads and assembled PacBio reads. We also illustrate the practical application on a Nanopore dataset that we sequenced from an anammox bio-rector community.ReviewersThis article was reviewed by Nicola Segata together with Moreno Zolfo, Pete James Lockhart and Serghei Mangul.ConclusionThis work extends the applicability of the widely-used metagenomic analysis software MEGAN to long reads. Our study suggests that the presented LAST+MEGAN-LR pipeline is sufficiently fast and accurate.

Accuracy of microbial community diversity estimated by closed- and open-reference OTUs.

Next-generation sequencing of 16S ribosomal RNA is widely used to survey microbial communities. Sequences are typically assigned to Operational Taxonomic Units (OTUs). Closed- and open-reference OTU assignment matches reads to a reference database at 97% identity (closed), then clusters unmatched reads using a de novo method (open). Implementations of these methods in the QIIME package were tested on several mock community datasets with 20 strains using different sequencing technologies and primers. Richness (number of reported OTUs) was often greatly exaggerated, with hundreds or thousands of OTUs generated on Illumina datasets. Between-sample diversity was also found to be highly exaggerated in many cases, with weighted Jaccard distances between identical mock samples often close to one, indicating very low similarity. Non-overlapping hyper-variable regions in 70% of species were assigned to different OTUs. On mock communities with Illumina V4 reads, 56% to 88% of predicted genus names were false positives. Biological inferences obtained using these methods are therefore not reliable.