2024

1. 

   Diagnosing Systematic Effects Using the Inferred Initial Power Spectrum

   Tristan Hoellinger, and Florent Leclercq

   arXiv e-prints Dec 2024

   Abs Bib HTML

   The next generation of galaxy surveys has the potential to substantially deepen our understanding of the Universe. This potential hinges on our ability to rigorously address systematic uncertainties. Until now, diagnosing systematic effects prior to inferring cosmological parameters has been out of reach in field-based implicit likelihood cosmological inference frameworks. As a solution, we aim to diagnose a variety of systematic effects in galaxy surveys prior to inferring cosmological parameters, using the inferred initial matter power spectrum. Our approach is built upon a two-step framework. First, we employ the Simulator Expansion for Likelihood-Free Inference (SELFI) algorithm to infer the initial matter power spectrum, which we utilise to thoroughly investigate the impact of systematic effects. This investigation relies on a single set of N-body simulations. Second, we obtain a posterior on cosmological parameters via implicit likelihood inference, recycling the simulations from the first step for data compression. For demonstration, we rely on a model of large-scale spectroscopic galaxy surveys that incorporates fully non-linear gravitational evolution and simulates multiple systematic effects encountered in real surveys. We provide a practical guide on how the SELFI posterior can be used to assess the impact of misspecified galaxy bias parameters, selection functions, survey masks, inaccurate redshifts, and approximate gravity models on the inferred initial matter power spectrum. We show that a subtly misspecified model can lead to a bias exceeding 2σin the (\Omega_\mathrmm,\sigma_8) plane, which we are able to detect and avoid prior to inferring the cosmological parameters. This framework has the potential to significantly enhance the robustness of physical information extraction from full-forward models of large-scale galaxy surveys such as DESI, Euclid, and LSST.

   @article{hoellinger2024,
     author = {{Hoellinger}, Tristan and {Leclercq}, Florent},
     title = {{Diagnosing Systematic Effects Using the Inferred Initial Power Spectrum}},
     journal = {arXiv e-prints},
     keywords = {Astrophysics - Cosmology and Nongalactic Astrophysics, Astrophysics - Instrumentation and Methods for Astrophysics},
     year = {2024},
     month = dec,
     eid = {arXiv:2412.04443},
     pages = {arXiv:2412.04443},
     archiveprefix = {arXiv},
     eprint = {2412.04443},
     primaryclass = {astro-ph.CO},
     adsurl = {https://ui.adsabs.harvard.edu/abs/2024arXiv241204443H},
     adsnote = {Provided by the SAO/NASA Astrophysics Data System},
   }

2023

1. 

   Enhancer/gene relationships: need for more reliable genome-wide reference sets

   Tristan Hoellinger, Camille Mestre, Hugues Aschard, and 4 more authors

   Frontiers in Bioinformatics 2023

   Abs Bib HTML

   Differences in cells’ functions arise from differential activity of regulatory elements, including enhancers. Enhancers are cis-regulatory elements that cooperate with promoters through transcription factors (TF) to activate the expression of one or several genes by getting physically close to them in the 3D space of the nucleus. There is increasing evidence that genetic variants associated with common diseases are enriched in enhancers active in cell types relevant to these diseases. Identifying the enhancers associated with genes and conversely, the sets of genes activated by each enhancer (the so-called enhancer/gene or E/G relationships) across cell types, can help understanding the genetic mechanisms underlying human diseases. There are three broad approaches for the genome-wide identification of E/G relationships in a cell type: (1) genetic link methods or eQTL, (2) functional link methods based on 1D functional data such as open chromatin, histone mark or gene expression and (3) spatial link methods based on 3D data such as HiC. Since (1) and (3) are costly, the current strategy is to develop functional link methods and to use data from (1) and (3) as reference to evaluate them. However, there is still no consensus on the best functional link method to date, and method comparison remain seldom. Here, we compared the relative performances of three recent methods for the identification of18 enhancer-gene links, TargetFinder, Average-Rank, and the ABC model, using the three19 latest benchmarks from the field: a reference that combines 3D and eQTL data, called BENGI,20 and two genetic screening references, called CRiFF and CRiSPRi. Overall, none of the three21 methods performed best on the three references. CRiFF and CRISPRi reference sets are likely22 more reliable, but CRiFF is not genome-wide and CRiFF and CRISPRi are mostly available on23 the K562 cancer cell line. The BENGI reference set is genome-wide but likely contains many false24 positives. This study therefore calls for new reliable and genome-wide E/G reference data rather25 than new functional link E/G identification methods.

   @article{hoellinger2023,
     title = {Enhancer/gene relationships: need for more reliable genome-wide reference sets},
     author = {Hoellinger, Tristan and Mestre, Camille and Aschard, Hugues and Le Goff, Wilfried and Foissac, Sylvain and Faraut, Thomas and Djebali, Sarah},
     journal = {Frontiers in Bioinformatics},
     year = {2023},
     volume = {3},
     section = {Insights in Bioinformatics},
     doi = {10.3389/fbinf.2023.1092853},
   }

2020

1. 

   Data-Driven Simulation for Augmented Surgery

   Andrea Mendizabal, Eleonora Tagliabue, Tristan Hoellinger, and 3 more authors

   2020

   Abs Bib HTML

   To build an augmented view of an organ during surgery, it is essential to have a biomechanical model with appropriate material parameters and boundary conditions, able to match patient specific properties. Adaptation to the patient’s anatomy is obtained by exploiting the image-rich context specific to our application domain. While information about the organ shape, for instance, can be obtained preoperatively, other patient-specific parameters can only be determined intraoperatively. To this end, we are developing data-driven simulations, which exploit information extracted from a stream of medical images. Such simulations need to run in realtime. To this end we have developed dedicated numerical methods, which allow for real-time computation of finite element simulations.

   @inbook{Mendizabal2020,
     author = {Mendizabal, Andrea and Tagliabue, Eleonora and Hoellinger, Tristan and Brunet, Jean-Nicolas and Nikolaev, Sergei and Cotin, St{\'e}phane},
     journal = {Advanced Structured Materials},
     editor = {Abali, Bilen Emek and Giorgio, Ivan},
     title = {Data-Driven Simulation for Augmented Surgery},
     booktitle = {Developments and Novel Approaches in Biomechanics and Metamaterials},
     year = {2020},
     publisher = {Springer International Publishing},
     address = {Cham},
     pages = {71--96},
     isbn = {978-3-030-50464-9},
     doi = {10.1007/978-3-030-50464-9_5},
     url = {https://doi.org/10.1007/978-3-030-50464-9_5},
   }