Showing 1–14 of 14 results for author: Eldar, A

Object detection under the linear subspace model with application to cryo-EM images

Authors: Amitay Eldar, Keren Mor Waknin, Samuel Davenport, Tamir Bendory, Armin Schwartzman, Yoel Shkolnisky

Abstract: Detecting multiple unknown objects in noisy data is a key problem in many scientific fields, such as electron microscopy imaging. A common model for the unknown objects is the linear subspace model, which assumes that the objects can be expanded in some known basis (such as the Fourier basis). In this paper, we develop an object detection algorithm that under the linear subspace model is asymptoti… ▽ More Detecting multiple unknown objects in noisy data is a key problem in many scientific fields, such as electron microscopy imaging. A common model for the unknown objects is the linear subspace model, which assumes that the objects can be expanded in some known basis (such as the Fourier basis). In this paper, we develop an object detection algorithm that under the linear subspace model is asymptotically guaranteed to detect all objects, while controlling the family wise error rate or the false discovery rate. Numerical simulations show that the algorithm also controls the error rate with high power in the non-asymptotic regime, even in highly challenging regimes. We apply the proposed algorithm to experimental electron microscopy data set, and show that it outperforms existing standard software. △ Less

Submitted 1 May, 2024; originally announced May 2024.

MSC Class: 6008; 60G15; 60G10; 60G35; 62M20; 62M40; ACM Class: G.3; I.4

Unsupervised particle sorting for cryo-EM using probabilistic PCA

Authors: Gili Weiss-Dicker, Amitay Eldar, Yoel Shkolinsky, Tamir Bendory

Abstract: Single-particle cryo-electron microscopy (cryo-EM) is a leading technology to resolve the structure of molecules. Early in the process, the user detects potential particle images in the raw data. Typically, there are many false detections as a result of high levels of noise and contamination. Currently, removing the false detections requires human intervention to sort the hundred thousands of imag… ▽ More Single-particle cryo-electron microscopy (cryo-EM) is a leading technology to resolve the structure of molecules. Early in the process, the user detects potential particle images in the raw data. Typically, there are many false detections as a result of high levels of noise and contamination. Currently, removing the false detections requires human intervention to sort the hundred thousands of images. We propose a statistically-established unsupervised algorithm to remove non-particle images. We model the particle images as a union of low-dimensional subspaces, assuming non-particle images are arbitrarily scattered in the high-dimensional space. The algorithm is based on an extension of the probabilistic PCA framework to robustly learn a non-linear model of union of subspaces. This provides a flexible model for cryo-EM data, and allows to automatically remove images that correspond to pure noise and contamination. Numerical experiments corroborate the effectiveness of the sorting algorithm. △ Less

Submitted 7 March, 2023; v1 submitted 23 October, 2022; originally announced October 2022.

ASOCEM: Automatic Segmentation Of Contaminations in cryo-EM

Authors: Amitay Eldar, Ido Amos, Yoel Shkolnisky

Abstract: Particle picking is currently a critical step in the cryo-electron microscopy single particle reconstruction pipeline. Contaminations in the acquired micrographs severely degrade the performance of particle pickers, resulting is many ``non-particles'' in the collected stack of particles. In this paper, we present ASOCEM (Automatic Segmentation Of Contaminations in cryo-EM), an automatic method to… ▽ More Particle picking is currently a critical step in the cryo-electron microscopy single particle reconstruction pipeline. Contaminations in the acquired micrographs severely degrade the performance of particle pickers, resulting is many ``non-particles'' in the collected stack of particles. In this paper, we present ASOCEM (Automatic Segmentation Of Contaminations in cryo-EM), an automatic method to detect and segment contaminations, which requires as an input only the approximated particle size. In particular, it does not require any parameter tuning nor manual intervention. Our method is based on the observation that the statistical distribution of contaminated regions is different from that of the rest of the micrograph. This nonrestrictive assumption allows to automatically detect various types of contaminations, from the carbon edges of the supporting grid to high contrast blobs of different sizes. We demonstrate the efficiency of our algorithm using various experimental data sets containing various types of contaminations. ASOCEM is integrated as part of the KLT picker \cite{ELDAR2020107473} and is available at \url{https://github.com/ShkolniskyLab/kltpicker2}. △ Less

Submitted 18 January, 2022; originally announced January 2022.

Biophysics underlying the swarm to biofilm transition

Authors: Vasco M. Worlitzer, Ajesh Jose, Ilana Grinberg, Markus Bär, Sebastian Heidenreich, Avigdor Eldar, Gil Ariel, Avraham Be'er

Abstract: Bacteria organize in a variety of collective states, from swarming, which has been attributed to rapid surface exploration, to biofilms, which are highly dense immobile communities attributed to stress resistance. It has been suggested that biofilm and swarming are oppositely controlled, making this transition particularly interesting for understanding the ability of bacterial colonies to adapt to… ▽ More Bacteria organize in a variety of collective states, from swarming, which has been attributed to rapid surface exploration, to biofilms, which are highly dense immobile communities attributed to stress resistance. It has been suggested that biofilm and swarming are oppositely controlled, making this transition particularly interesting for understanding the ability of bacterial colonies to adapt to challenging environments. Here, the swarm to biofilm transition is studied experimentally by analyzing the bacterial dynamics both on the individual and collective scales. We show that both biological and physical processes facilitate the transition - a few individual cells that initiate the biofilm program cause nucleation of large, scale-free stationary aggregates of trapped swarm cells. Around aggregates, cells continue swarming almost unobstructed, while inside, trapped cells slowly transform to biofilm. While our experimental findings rule out previously suggested purely physical effects as a trigger for biofilm formation, they show how physical processes, such as clustering and jamming, accelerate biofilm formation. △ Less

Submitted 21 October, 2021; originally announced October 2021.

Journal ref: Science Advances 2022

KLT Picker: Particle Picking Using Data-Driven Optimal Templates

Authors: Amitay Eldar, Boris Landa, Yoel Shkolnisky

Abstract: Particle picking is currently a critical step in the cryo-EM single particle reconstruction pipeline. Despite extensive work on this problem, for many data sets it is still challenging, especially for low SNR micrographs. We present the KLT (Karhunen Loeve Transform) picker, which is fully automatic and requires as an input only the approximated particle size. In particular, it does not require an… ▽ More Particle picking is currently a critical step in the cryo-EM single particle reconstruction pipeline. Despite extensive work on this problem, for many data sets it is still challenging, especially for low SNR micrographs. We present the KLT (Karhunen Loeve Transform) picker, which is fully automatic and requires as an input only the approximated particle size. In particular, it does not require any manual picking. Our method is designed especially to handle low SNR micrographs. It is based on learning a set of optimal templates through the use of multi-variate statistical analysis via the Karhunen Loeve Transform. We evaluate the KLT picker on publicly available data sets and present high-quality results with minimal manual effort. △ Less

Submitted 12 December, 2019; originally announced December 2019.

Simulating the Formation of the Local Galaxy Population

Authors: H. Mathis, G. Lemson, V. Springel, G. Kauffmann, S. D. M. White, A. Eldar, A. Dekel

Abstract: We simulate the formation and evolution of the local galaxy population starting from initial conditions with a smoothed linear density field which matches that derived from the IRAS 1.2 Jy galaxy survey. Our simulations track the formation and evolution of all dark matter haloes more massive than 10e+11 solar masses out to a distance of 8000 km/s from the Milky Way. We implement prescriptions si… ▽ More We simulate the formation and evolution of the local galaxy population starting from initial conditions with a smoothed linear density field which matches that derived from the IRAS 1.2 Jy galaxy survey. Our simulations track the formation and evolution of all dark matter haloes more massive than 10e+11 solar masses out to a distance of 8000 km/s from the Milky Way. We implement prescriptions similar to those of Kauffmann et al. (1999) to follow the assembly and evolution of the galaxies within these haloes. We focus on two variants of the CDM cosmology: an LCDM and a tCDM model. Galaxy formation in each is adjusted to reproduce the I-band Tully-Fisher relation of Giovanelli et al. (1997). We compare the present-day luminosity functions, colours, morphology and spatial distribution of our simulated galaxies with those of the real local population, in particular with the Updated Zwicky Catalog, with the IRAS PSCz redshift survey, and with individual local clusters such as Coma, Virgo and Perseus. We also use the simulations to study the clustering bias between the dark matter and galaxies of differing type. Although some significant discrepancies remain, our simulations recover the observed intrinsic properties and the observed spatial distribution of local galaxies reasonably well. They can thus be used to calibrate methods which use the observed local galaxy population to estimate the cosmic density parameter or to draw conclusions about the mechanisms of galaxy formation. To facilitate such work, we publically release our z=0 galaxy catalogues, together with the underlying mass distribution. △ Less

Submitted 5 November, 2001; originally announced November 2001.

Comments: 25 pages, 20 figures, submitted to MNRAS. High resolution copies of figures 1 and 3, halo and galaxy catalogues can be found at http://www.mpa-garching.mpg.de/NumCos/CR/index.html

Journal ref: Mon.Not.Roy.Astron.Soc. 333 (2002) 739

Peculiar Velocity Reconstruction with Fast Action Method: Tests on Mock Redshift Surveys

Authors: Enzo Branchini, Amiram Eldar, Adi Nusser

Abstract: We present extensive tests of the Fast Action Method (FAM) for recovering the past orbits of mass tracers in an expanding universe from their redshift-space coordinates at the present epoch. The tests focus on the reconstruction of present-day peculiar velocities using mock catalogs extracted from high resolution $N$-body simulations. The method allows for a self-consistent treatment of redshift… ▽ More We present extensive tests of the Fast Action Method (FAM) for recovering the past orbits of mass tracers in an expanding universe from their redshift-space coordinates at the present epoch. The tests focus on the reconstruction of present-day peculiar velocities using mock catalogs extracted from high resolution $N$-body simulations. The method allows for a self-consistent treatment of redshift-space distortions by direct minimization of a modified action for a cosmological gravitating system. When applied to ideal, volume limited catalogs, FAM recovers unbiased peculiar velocities with a 1-D, 1σerror of ~220 km/s, if velocities are smoothed on a scale of 5 Mpc/h. Alternatively, when no smoothing is applied, FAM predicts nearly unbiased velocities for objects residing outside the highest density regions. In this second case the 1σ$error decreases to a level of ~150 km/s. The correlation properties of the peculiar velocity fields are also correctly recovered on scales larger than 5 Mpc/h. Similar results are obtained when FAM is applied to flux limited catalogs mimicking the IRAS PSCz survey. In this case FAM reconstructs peculiar velocities with similar intrinsic random errors, while velocity-velocity correlation properties are well reproduced beyond scales of ~8 Mpc/h. We also show that FAM provides better velocity predictions than other, competing methods based on linear theory or Zel'dovich approximation. These results indicate that FAM can be successfully applied to presently available galaxy redshift surveys such as IRAS PSCz. △ Less

Submitted 3 May, 2002; v1 submitted 29 October, 2001; originally announced October 2001.

Comments: 26 pages, 16 figures. Figures 1,2,3,4,5,7,11,12 and 16 are also included as separate gif files. Added 2 new sections(5.3 and 6.3) and figures (11 and 16). More discussion added to section 7 (Summary and Conclusions). MNRAS accepted

Journal ref: Mon.Not.Roy.Astron.Soc. 335 (2002) 53

The Large-Scale Tidal Velocity Field

Authors: Y. Hoffman, A. Eldar, S. Zaroubi, A. Dekel

Abstract: We present a method for decomposing the cosmological velocity field in a given volume into its divergent component due to the density fluctuations inside the volume, and its tidal component due to the matter distribution outside the volume. The input consists of the density and velocity fields that are reconstructed either by POTENT or by Wiener Filter from a survey of peculiar velocities. The t… ▽ More We present a method for decomposing the cosmological velocity field in a given volume into its divergent component due to the density fluctuations inside the volume, and its tidal component due to the matter distribution outside the volume. The input consists of the density and velocity fields that are reconstructed either by POTENT or by Wiener Filter from a survey of peculiar velocities. The tidal field is further decomposed into a bulk velocity and a shear field. The method is applied here to the Mark III data within a sphere of radius 60 Mpc/h about the Local Group, and to the SFI data for comparison. We find that the tidal field contributes about half of the Local-Group velocity with respect to the CMB, with the tidal bulk velocity pointing to within ~ 30 degrees of the CMB dipole. The eigenvector with the largest eigenvalue of the shear tensor is aligned with the tidal bulk velocity to within ~ 40 degrees. The tidal field thus indicates the important dynamical role of a super attractor of mass (2-5) x 10^17 M_sun/h Omega^0.4 at ~ 150 Mpc/h, coinciding with the Shapley Concentration. There is also a hint for the dynamical role of two big voids in the Supergalactic Plane. The results are consistent for the two data sets and the two methods of reconstruction. △ Less

Submitted 11 February, 2001; originally announced February 2001.

Comments: 21 pages, LaTex, uses aaspp4.sty, submitted to ApJ

Nonlinear Peculiar-Velocity Analysis and PCA

Authors: A. Dekel, A. Eldar, L. Silberman, I. Zehavi

Abstract: We allow for nonlinear effects in the likelihood analysis of peculiar velocities, and obtain ~35%-lower values for the cosmological density parameter and for the amplitude of mass-density fluctuations. The power spectrum in the linear regime is assumed to be of the flat LCDM model (h=0.65, n=1) with only Om_m free. Since the likelihood is driven by the nonlinear regime, we "break" the power spec… ▽ More We allow for nonlinear effects in the likelihood analysis of peculiar velocities, and obtain ~35%-lower values for the cosmological density parameter and for the amplitude of mass-density fluctuations. The power spectrum in the linear regime is assumed to be of the flat LCDM model (h=0.65, n=1) with only Om_m free. Since the likelihood is driven by the nonlinear regime, we "break" the power spectrum at k_b=0.2 h/Mpc and fit a two-parameter power-law at k>k_b. This allows for an unbiased fit in the linear regime. Tests using improved mock catalogs demonstrate a reduced bias and a better fit. We find for the Mark III and SFI data Om_m=0.35+-0.09$ with sigma_8*Om_m^0.6=0.55+-0.10 (90% errors). When allowing deviations from \lcdm, we find an indication for a wiggle in the power spectrum in the form of an excess near k~0.05 and a deficiency at k~0.1 h/Mpc --- a "cold flow" which may be related to a feature indicated from redshift surveys and the second peak in the CMB anisotropy. A chi^2 test applied to principal modes demonstrates that the nonlinear procedure improves the goodness of fit. The Principal Component Analysis (PCA) helps identifying spatial features of the data and fine-tuning the theoretical and error models. We address the potential for optimal data compression using PCA. △ Less

Submitted 28 January, 2001; originally announced January 2001.

Comments: 15 pages, LaTex, in Mining the Sky, July 31 - August 4, 2000, Garching, Germany

Cosmological Density and Power Spectrum from Peculiar Velocities: Nonlinear Corrections and PCA

Authors: L. Silberman, A. Dekel, A. Eldar, I. Zehavi

Abstract: We allow for nonlinear effects in the likelihood analysis of galaxy peculiar velocities, and obtain ~35%-lower values for the cosmological density parameter Om and the amplitude of mass-density fluctuations. The power spectrum in the linear regime is assumed to be a flat LCDM model (h=0.65, n=1, COBE) with only Om as a free parameter. Since the likelihood is driven by the nonlinear regime, we "b… ▽ More We allow for nonlinear effects in the likelihood analysis of galaxy peculiar velocities, and obtain ~35%-lower values for the cosmological density parameter Om and the amplitude of mass-density fluctuations. The power spectrum in the linear regime is assumed to be a flat LCDM model (h=0.65, n=1, COBE) with only Om as a free parameter. Since the likelihood is driven by the nonlinear regime, we "break" the power spectrum at k_b=0.2 h/Mpc and fit a power law at k>k_b. This allows for independent matching of the nonlinear behavior and an unbiased fit in the linear regime. The analysis assumes Gaussian fluctuations and errors, and a linear relation between velocity and density. Tests using proper mock catalogs demonstrate a reduced bias and a better fit. We find for the Mark3 and SFI data Om_m=0.32+-0.06 and 0.37+-0.09 respectively, with sigma_8*Om^0.6 = 0.49+-0.06 and 0.63+-0.08, in agreement with constraints from other data. The quoted 90% errors include cosmic variance. The improvement in likelihood due to the nonlinear correction is very significant for Mark3 and moderately so for SFI. When allowing deviations from LCDM, we find an indication for a wiggle in the power spectrum: an excess near k=0.05 and a deficiency at k=0.1 (cold flow). This may be related to the wiggle seen in the power spectrum from redshift surveys and the second peak in the CMB anisotropy. A chi^2 test applied to modes of a Principal Component Analysis (PCA) shows that the nonlinear procedure improves the goodness of fit and reduces a spatial gradient of concern in the linear analysis. The PCA allows addressing spatial features of the data and fine-tuning the theoretical and error models. It shows that the models used are appropriate for the cosmological parameter estimation performed. We address the potential for optimal data compression using PCA. △ Less

Submitted 24 April, 2001; v1 submitted 21 January, 2001; originally announced January 2001.

Comments: 18 pages, LaTex, uses emulateapj.sty, ApJ in press (August 10, 2001), improvements to text and figures, updated references

Journal ref: Astrophys.J. 557 (2001) 102-116

On the Viewing Angle Dependence of Blazar Variability

Authors: Avigdor Eldar, Amir Levinson

Abstract: Internal shocks propagating through an ambient radiation field, are subject to a radiative drag that, under certain conditions, can significantly affect their dynamics and, consequently, the evolution of the beaming cone of emission produced behind the shocks. The resultant change of the Doppler factor combined with opacity effects leads to a strong dependence of the variability pattern produced… ▽ More Internal shocks propagating through an ambient radiation field, are subject to a radiative drag that, under certain conditions, can significantly affect their dynamics and, consequently, the evolution of the beaming cone of emission produced behind the shocks. The resultant change of the Doppler factor combined with opacity effects leads to a strong dependence of the variability pattern produced by such systems, specifically, the shape of the light curves and the characteristics of correlated emission, on viewing angle. One implication is that objects oriented at relatively large viewing angles to the observer should exhibit a higher level of activity at high synchrotron frequencies (above the self-absorption frequency) and at gamma-ray energies below the threshold energy to pair production, than at lower (radio/millimeter) frequencies. △ Less

Submitted 20 January, 2000; originally announced January 2000.

Comments: 6 pages, accepted by MNRAS

Large-Scale Power Spectrum and Cosmological Parameters from SFI Peculiar Velocities

Authors: Wolfram Freudling, Idit Zehavi, Luiz N. da Costa, Avishai Dekel, Amiram Eldar, Riccardo Giovanelli, Martha P. Haynes, John J. Salzer, Gary Wegner, Saleem Zaroubi

Abstract: We estimate the power spectrum of mass density fluctuations from peculiar velocities of galaxies by applying an improved maximum-likelihood technique to the new all-sky SFI catalog. Parametric models are used for the power spectrum and the errors, and the free parameters are determined by assuming Gaussian velocity fields and errors and maximizing the probability of the data given the model. It… ▽ More We estimate the power spectrum of mass density fluctuations from peculiar velocities of galaxies by applying an improved maximum-likelihood technique to the new all-sky SFI catalog. Parametric models are used for the power spectrum and the errors, and the free parameters are determined by assuming Gaussian velocity fields and errors and maximizing the probability of the data given the model. It has been applied to generalized CDM models with and without COBE normalization. The method has been carefully tested using artificial SFI catalogs. The most likely distance errors are found to be similar to the original error estimates in the SFI data. The general result that is not very sensitive to the prior model used is a relatively high amplitude of the power spectrum. For example, at k=0.1 h/Mpc we find P(k)Ω^{1.2}=(4.4+/-1.7)X10^3 (Mpc/h)^3. An integral over the power spectrum yields σ_8Ω^{0.6}=0.82+/-0.12. Model-dependent constraints on the cosmological parameters are obtained for families of CDM models. For example, for COBE-normalized ΛCDM models (scalar fluctuations only), the maximum-likelihood result can be approximated by Ωn^{2} h_{60}^{1.3} =0.58+/-0.11. The formal random errors quoted correspond to the 90% confidence level. The total uncertainty, including systematic errors associated with nonlinear effects, may be larger by a factor of ~2. These results are in agreement with an application of a similar method to other data (Mark III). △ Less

Submitted 9 April, 1999; originally announced April 1999.

Comments: 35 pages, 11 figures. To appear in the Astrophysical Journal

Report number: FERMILAB-Pub-99/079-A

Journal ref: Astrophys.J. 523 (1999) 1

POTENT Reconstruction from Mark III Velocities

Authors: A. Dekel, A. Eldar, T. Kolatt, A. Yahil, J. A. Willick, S. M. Faber, S. Courteau, D. Burstein

Abstract: We present an improved POTENT method for reconstructing the velocity and mass density fields from radial peculiar velocities, test it with mock catalogs, and apply it to the Mark III Catalog. Method improvments: (a) inhomogeneous Malmquist bias is reduced by grouping and corrected in forward or inverse analyses of inferred distances, (b) the smoothing into a radial velocity field is optimized to… ▽ More We present an improved POTENT method for reconstructing the velocity and mass density fields from radial peculiar velocities, test it with mock catalogs, and apply it to the Mark III Catalog. Method improvments: (a) inhomogeneous Malmquist bias is reduced by grouping and corrected in forward or inverse analyses of inferred distances, (b) the smoothing into a radial velocity field is optimized to reduce window and sampling biases, (c) the density is derived from the velocity using an improved nonlinear approximation, and (d) the computational errors are made negligible. The method is tested and optimized using mock catalogs based on an N-body simulation that mimics our cosmological neighborhood, and the remaining errors are evaluated quantitatively. The Mark III catalog, with ~3300 grouped galaxies, allows a reliable reconstruction with fixed Gaussian smoothing of 10-12 Mpc/h out to ~60 Mpc/h. We present maps of the 3D velocity and mass-density fields and the corresponding errors. The typical systematic and random errors in the density fluctuations inside 40 Mpc/h are \pm 0.13 and \pm 0.18. The recovered mass distribution resembles in its gross features the galaxy distribution in redshift surveys and the mass distribution in a similar POTENT analysis of a complementary velocity catalog (SFI), including the Great Attractor, Perseus-Pisces, and the void in between. The reconstruction inside ~40 Mpc/h is not affected much by a revised calibration of the distance indicators (VM2, tailored to match the velocities from the IRAS 1.2Jy redshift survey). The bulk velocity within the sphere of radius 50 Mpc/h about the Local Group is V_50=370 \pm 110 km/s (including systematic errors), and is shown to be mostly generated by external mass fluctuations. With the VM2 calibration, V_50 is reduced to 305 \pm 110 km/s. △ Less

Submitted 9 December, 1998; originally announced December 1998.

Comments: 60 pages, LaTeX, 3 tables and 27 figures incorporated (may print the most crucial figures only, by commenting out one line in the LaTex source)

IRAS versus POTENT Density Fields on Large Scales: Biasing and Omega

Authors: Y. Sigad, A. Eldar, A. Dekel, M. A. Strauss, A. Yahil

Abstract: The galaxy density field as extracted from the IRAS 1.2 Jy redshift survey is compared to the mass density field as reconstructed by the POTENT method from the Mark III catalog of peculiar velocities. The reconstruction is done with Gaussian smoothing of radius 12 h^{-1}Mpc, and the comparison is carried out within volumes of effective radii 31-46 h^{-1}Mpc, containing approximately 10-26 indepe… ▽ More The galaxy density field as extracted from the IRAS 1.2 Jy redshift survey is compared to the mass density field as reconstructed by the POTENT method from the Mark III catalog of peculiar velocities. The reconstruction is done with Gaussian smoothing of radius 12 h^{-1}Mpc, and the comparison is carried out within volumes of effective radii 31-46 h^{-1}Mpc, containing approximately 10-26 independent samples. Random and systematic errors are estimated from multiple realizations of mock catalogs drawn from a simulation that mimics the observed density field in the local universe. The relationship between the two density fields is found to be consistent with gravitational instability theory in the mildly nonlinear regime and a linear biasing relation between galaxies and mass. We measure beta = Omega^{0.6}/b_I = 0.89 \pm 0.12 within a volume of effective radius 40 h^{-1}Mpc, where b_I is the IRAS galaxy biasing parameter at 12 h^{-1}Mpc. This result is only weakly dependent on the comparison volume, suggesting that cosmic scatter is no greater than \pm 0.1. These data are thus consistent with Omega=1 and b_I\approx 1. If b_I>0.75, as theoretical models of biasing indicate, then Omega>0.33 at 95% confidence. A comparison with other estimates of beta suggests scale-dependence in the biasing relation for IRAS galaxies. △ Less

Submitted 13 August, 1997; originally announced August 1997.

Comments: 35 pages including 10 figures, AAS Latex, Submitted to The Astrophysical Journal