-
Neural Networks beyond explainability: Selective inference for sequence motifs
Authors:
Antoine Villié,
Philippe Veber,
Yohann de Castro,
Laurent Jacob
Abstract:
Over the past decade, neural networks have been successful at making predictions from biological sequences, especially in the context of regulatory genomics. As in other fields of deep learning, tools have been devised to extract features such as sequence motifs that can explain the predictions made by a trained network. Here we intend to go beyond explainable machine learning and introduce SEISM,…
▽ More
Over the past decade, neural networks have been successful at making predictions from biological sequences, especially in the context of regulatory genomics. As in other fields of deep learning, tools have been devised to extract features such as sequence motifs that can explain the predictions made by a trained network. Here we intend to go beyond explainable machine learning and introduce SEISM, a selective inference procedure to test the association between these extracted features and the predicted phenotype. In particular, we discuss how training a one-layer convolutional network is formally equivalent to selecting motifs maximizing some association score. We adapt existing sampling-based selective inference procedures by quantizing this selection over an infinite set to a large but finite grid. Finally, we show that sampling under a specific choice of parameters is sufficient to characterize the composite null hypothesis typically used for selective inference-a result that goes well beyond our particular framework. We illustrate the behavior of our method in terms of calibration, power and speed and discuss its power/speed trade-off with a simpler data-split strategy. SEISM paves the way to an easier analysis of neural networks used in regulatory genomics, and to more powerful methods for genome wide association studies (GWAS).
△ Less
Submitted 23 December, 2022;
originally announced December 2022.
-
Stability of the tetragonal phase of BaZrO3 under high pressure
Authors:
Constance Toulouse,
Danila Amoroso,
Robert Oliva,
Cong Xin,
Pierre Bouvier,
Pierre Fertey,
Philippe Veber,
Mario Maglione,
Philippe Ghosez,
Jens Kreisel,
Mael Guennou
Abstract:
In this paper, we revisit the high-pressure behavior of BaZrO3 by a combination of first-principle calculations, Raman spectroscopy, and x-ray diffraction under high pressure. We confirm experimentally the cubic-to-tetragonal transition at 10 GPa and find no evidence for any other phase transition up to 45 GPa, the highest pressures investigated, at variance with past reports. We re-investigate ph…
▽ More
In this paper, we revisit the high-pressure behavior of BaZrO3 by a combination of first-principle calculations, Raman spectroscopy, and x-ray diffraction under high pressure. We confirm experimentally the cubic-to-tetragonal transition at 10 GPa and find no evidence for any other phase transition up to 45 GPa, the highest pressures investigated, at variance with past reports. We re-investigate phase stability with density functional theory considering not only the known tetragonal (I4/mcm) phase but also other potential antiferrodistortive candidates. This shows that the tetragonal phase becomes progressively more stable upon increasing pressure as compared to phases with more complex tilt systems. The possibility for a second transition to another tilted phase at higher pressures, and in particular to the very common orthorhombic Pnma structure, is therefore ruled out.
△ Less
Submitted 16 August, 2022; v1 submitted 13 May, 2022;
originally announced May 2022.
-
Structural and lattice-dynamical properties of Tb2O3 under compression: a comparative study with rare-earth and related sesquioxides
Authors:
Jordi Ibáñez,
Juan Ángel Sans,
Vanesa Cuenca-Gotor,
Robert Oliva,
Óscar Gomis,
Plácida Rodríguez-Hernández,
Alfonso Muñoz,
Ulises Rodríguez-Mendoza,
Matías Velázquez,
Philippe Veber,
Catalin Popescu,
Francisco Javier Manjón
Abstract:
We report a joint experimental and theoretical investigation of the high-pressure structural and vibrational properties of terbium sesquioxide (Tb2O3). Powder x-ray diffraction and Raman scattering measurements show that cubic (C-type) Tb2O3 undergoes two phase transitions up to 25 GPa. We observe a first irreversible reconstructive transition to the monoclinic (B-type) phase at ~7 GPa and a subse…
▽ More
We report a joint experimental and theoretical investigation of the high-pressure structural and vibrational properties of terbium sesquioxide (Tb2O3). Powder x-ray diffraction and Raman scattering measurements show that cubic (C-type) Tb2O3 undergoes two phase transitions up to 25 GPa. We observe a first irreversible reconstructive transition to the monoclinic (B-type) phase at ~7 GPa and a subsequent reversible displacive transition from the monoclinic to the trigonal (A-type) phase at ~12 GPa. Thus, Tb2O3 is found to follow the well-known C-B-A phase transition found in other cubic rare-earth sesquioxides with cations of larger atomic mass than Tb. Our ab initio theoretical calculations predict phase transition pressures and bulk moduli for the three phases in rather good agreement with experimental results. Moreover, Raman-active modes of the three phases have been monitored as a function of pressure and lattice-dynamics calculations have allowed us to confirm the assignment of the experimental phonon modes in the C-type phase as well as to make a tentative assignment of the symmetry of most vibrational modes in the B- and A-type phases. Finally, we extract the bulk moduli and the Raman-active modes frequencies together with their pressure coefficients for the three phases of Tb2O3. These results are thoroughly compared and discussed in relation to those reported for rare-earth and other related sesquioxides. It is concluded that the evolution of the volume and bulk modulus of all the three phases of these technologically relevant compounds exhibit a nearly linear trend with respect to the third power of the ionic radii of the cations.
△ Less
Submitted 26 May, 2020;
originally announced May 2020.
-
Lattice dynamics and Raman spectrum of BaZrO3 single crystals
Authors:
Constance Toulouse,
Danila Amoroso,
Cong Xin,
Philippe Veber,
Monica Ciomaga Hatnean,
Geetha Balakrishnan,
Mario Maglione,
Philippe Ghosez,
Jens Kreisel,
Mael Guennou
Abstract:
BaZrO3 is a perovskite that remains in the simple cubic phase at all temperatures, hence with no first-order Raman-active phonon mode allowed by symmetry. Yet, it exhibits an intense Raman spectrum with sharp and well-defined features. Here, we report the evolution of the Raman spectrum of BaZrO3 single crystals in a broad temperature range (4--1200~K) and discuss its origin with the support of de…
▽ More
BaZrO3 is a perovskite that remains in the simple cubic phase at all temperatures, hence with no first-order Raman-active phonon mode allowed by symmetry. Yet, it exhibits an intense Raman spectrum with sharp and well-defined features. Here, we report the evolution of the Raman spectrum of BaZrO3 single crystals in a broad temperature range (4--1200~K) and discuss its origin with the support of detailed first-principle calculations of the lattice dynamics. Phonon calculations are performed not only for the cubic phase of BaZrO3, but also for the low-symmetry phases with octahedra tilts that have been suspected to exist at the nanoscale. We show that the Raman spectrum shows no direct evidence for these nanodomains, but can instead be explained by classical second-order Raman scattering. We provide an assignment of the dominant features to phonon modes combinations. In particular, we show that the high frequency range of the spectrum is dominated by overtones and shows an image of the phonon density of states corresponding to the stretching modes of the oxygen octahedra.
△ Less
Submitted 3 July, 2019;
originally announced July 2019.
-
First scintillating bolometer tests of a CLYMENE R&D on Li$_2$MoO$_4$ scintillators towards a large-scale double-beta decay experiment
Authors:
G. Buşe,
A. Giuliani,
P. de Marcillac,
S. Marnieros,
C. Nones,
V. Novati,
E. Olivieri,
D. V. Poda,
T. Redon,
J. -B. Sand,
P. Veber,
M. Velázquez,
A. S. Zolotarova
Abstract:
A new R&D on lithium molybdate scintillators has begun within a project CLYMENE (Czochralski growth of Li$_2$MoO$_4$ crYstals for the scintillating boloMeters used in the rare EveNts sEarches). One of the main goals of the CLYMENE is a realization of a Li$_2$MoO$_4$ crystal growth line to be complementary to the one recently developed by LUMINEU in view of a mass production capacity for CUPID, a n…
▽ More
A new R&D on lithium molybdate scintillators has begun within a project CLYMENE (Czochralski growth of Li$_2$MoO$_4$ crYstals for the scintillating boloMeters used in the rare EveNts sEarches). One of the main goals of the CLYMENE is a realization of a Li$_2$MoO$_4$ crystal growth line to be complementary to the one recently developed by LUMINEU in view of a mass production capacity for CUPID, a next-generation tonne-scale bolometric experiment to search for neutrinoless double-beta decay. In the present paper we report the investigation of performance and radiopurity of 158-g and 13.5-g scintillating bolometers based on a first large-mass (230 g) Li$_2$MoO$_4$ crystal scintillator developed within the CLYMENE project. In particular, a good energy resolution (2--7 keV FWHM in the energy range of 0.2--5 MeV), one of the highest light yield (0.97 keV/MeV) amongst Li$_2$MoO$_4$ scintillating bolometers, an efficient alpha particles discrimination (10$σ$) and potentially low internal radioactive contamination (below 0.2--0.3 mBq/kg of U/Th, but 1.4 mBq/kg of $^{210}$Po) demonstrate prospects of the CLYMENE in the development of high quality and radiopure Li$_2$MoO$_4$ scintillators for CUPID.
△ Less
Submitted 8 March, 2018; v1 submitted 24 January, 2018;
originally announced January 2018.
-
MOSAIC_SSD: a new web-tool for the Species Sensitivity Distribution, allowing to include censored data by maximum likelihood
Authors:
Guillaume Kon Kam King,
Philippe Veber,
Sandrine Charles,
Marie Laure Delignette-Muller
Abstract:
Censored data are seldom taken into account in Species Sensitivity Distribution (SSD) analysis. However, they are found in virtually every dataset and sometimes represent the better part of the data. Stringent recommendations on data quality often lead to discard a lot of this meaningful data, often resulting in datasets of reduced size, which lack representativeness of any realistic community. Ho…
▽ More
Censored data are seldom taken into account in Species Sensitivity Distribution (SSD) analysis. However, they are found in virtually every dataset and sometimes represent the better part of the data. Stringent recommendations on data quality often lead to discard a lot of this meaningful data, often resulting in datasets of reduced size, which lack representativeness of any realistic community. However, it is reasonably simple to include censored data into SSD by using an extension of the standard maximum likelihood method. In this paper, we detail this approach based on the use of the R-package \emph{fitdistrplus}, dedicated to the fit of parametric probability distributions. In particular, we introduce the new web-tool MOSAIC$\_$SSD to fit an SSD on datasets containing any kind of data, censored or not. MOSAIC$\_$SSD allows predicting any Hazardous Concentration (HC) and provides in addition bootstrap confidence intervals on the prediction. In the end, taking examples from published data, we illustrate the added value of including censored data in SSD.
△ Less
Submitted 3 February, 2014; v1 submitted 22 November, 2013;
originally announced November 2013.
-
Statistical handling of reproduction data for exposure-response modelling
Authors:
Marie Laure Delignette-Muller,
Christelle Lopes,
Philippe Veber,
Sandrine Charles
Abstract:
Reproduction data collected through standard bioassays are classically analyzed by regression in order to fit exposure-response curves and estimate ECx values (x% Effective Concentration). But regression is often misused on such data, ignoring statistical issues related to i) the special nature of reproduction data (count data), ii) a potential inter-replicate variability and iii) a possible conco…
▽ More
Reproduction data collected through standard bioassays are classically analyzed by regression in order to fit exposure-response curves and estimate ECx values (x% Effective Concentration). But regression is often misused on such data, ignoring statistical issues related to i) the special nature of reproduction data (count data), ii) a potential inter-replicate variability and iii) a possible concomitant mortality. This paper offers new insights in dealing with those issues. Concerning mortality, particular attention was paid not to waste any valuable data - by drop** all the replicates with mortality - or to bias ECx values. For that purpose we defined a new covariate summing the observation periods during which each individual contributes to the reproduction process. This covariate was then used to quantify reproduction - for each replicate at each concentration - as a number of offspring per individual-day. We formulated three exposure-response models differing by their stochastic part. Those models were fitted to four datasets and compared using a Bayesian framework. The individual-day unit proved to be a suitable approach to use all the available data and prevent bias in the estimation of ECx values. Furthermore, a non-classical negative-binomial model was shown to correctly describe the inter-replicate variability observed in the studied datasets.
△ Less
Submitted 26 April, 2014; v1 submitted 10 October, 2013;
originally announced October 2013.
-
Detecting Inconsistencies in Large Biological Networks with Answer Set Programming
Authors:
Martin Gebser,
Torsten Schaub,
Sven Thiele,
Philippe Veber
Abstract:
We introduce an approach to detecting inconsistencies in large biological networks by using Answer Set Programming (ASP). To this end, we build upon a recently proposed notion of consistency between biochemical/genetic reactions and high-throughput profiles of cell activity. We then present an approach based on ASP to check the consistency of large-scale data sets. Moreover, we extend this methodo…
▽ More
We introduce an approach to detecting inconsistencies in large biological networks by using Answer Set Programming (ASP). To this end, we build upon a recently proposed notion of consistency between biochemical/genetic reactions and high-throughput profiles of cell activity. We then present an approach based on ASP to check the consistency of large-scale data sets. Moreover, we extend this methodology to provide explanations for inconsistencies by determining minimal representations of conflicts. In practice, this can be used to identify unreliable data or to indicate missing reactions.
△ Less
Submitted 1 July, 2010;
originally announced July 2010.
-
Lagrangian Approaches for a class of Matching Problems in Computational Biology
Authors:
Nicola Yanev,
Rumen Andonov,
Philippe Veber,
Stefan Balev
Abstract:
This paper presents efficient algorithms for solving the problem of aligning a protein structure template to a query amino-acid sequence, known as protein threading problem. We consider the problem as a special case of graph matching problem. We give formal graph and integer programming models of the problem. After studying the properties of these models, we propose two kinds of Lagrangian relax…
▽ More
This paper presents efficient algorithms for solving the problem of aligning a protein structure template to a query amino-acid sequence, known as protein threading problem. We consider the problem as a special case of graph matching problem. We give formal graph and integer programming models of the problem. After studying the properties of these models, we propose two kinds of Lagrangian relaxation for solving them. We present experimental results on real life instances showing the efficiency of our approaches.
△ Less
Submitted 1 September, 2006;
originally announced September 2006.
-
Complex Qualitative Models in Biology: a new approach
Authors:
Philippe Veber,
Michel Le_Borgne,
Anne Siegel,
Sandrine Lagarrigue,
Ovidiu Radulescu
Abstract:
We advocate the use of qualitative models in the analysis of large biological systems. We show how qualitative models are linked to theoretical differential models and practical graphical models of biological networks. A new technique for analyzing qualitative models is introduced, which is based on an efficient representation of qualitative systems. As shown through several applications, this r…
▽ More
We advocate the use of qualitative models in the analysis of large biological systems. We show how qualitative models are linked to theoretical differential models and practical graphical models of biological networks. A new technique for analyzing qualitative models is introduced, which is based on an efficient representation of qualitative systems. As shown through several applications, this representation is a relevant tool for the understanding and testing of large and complex biological networks.
△ Less
Submitted 13 April, 2006;
originally announced April 2006.
