Alexandre Bouchard-Côté

Annealed sequential Monte Carlo (ASMC) samplers are special cases of SMC samplers where the sequence of distributions can be embedded in a smooth path of distributions. Using this underlying path and a performance model based on the variance of the normalizing constant estimator, we systematically study dense-schedule limits. From our theory emerges a notion of global barrier, capturing the inherent complexity of normalizing constant approximation under our performance model. We then turn the resulting approximations into surrogate objective functions of algorithm performance, using them to guide method development. This leads to novel adaptive methods, optimized annealed SMC (OASMC), which address practical difficulties inherent in previous adaptive SMC methods. First, our OASMC algorithms are predictable: they produce a sequence of increasingly precise estimates at deterministic, known times. Second, optimized annealed importance sampling (OAIS), a special case of OASMC, enables schedule adaptation at a memory cost constant in the number of particles, requiring significantly less communication. Finally, these characteristics make OAIS highly efficient on GPUs. We provide an open-source, high-performance GPU implementation of our method and demonstrate up to a hundred-fold speed improvement compared to state-of-the-art adaptive AIS methods.

Gaia Data Release 4 epoch astrometry will enable the detection of thousands of exoplanets through astrometric motion. We present a composite catalog and modeling framework that extracts the maximum information from existing Hipparcos and Gaia data releases. We calibrate Gaia DR2 proper motions and DR3–DR2 scaled position differences against the Gaia DR3 reference frame, and combine these with the Hipparcos-Gaia Catalog of Accelerations (or HGCA), the Hipparcos intermediate astrometric data, Gaia astrometric excess noise, and Gaia radial velocity (RV) variability constraints. We implement a joint likelihood model for these data in Octofitter that marginalizes over Gaia’s unpublished observation epochs. This results in full orbit posteriors that can be computed uniformly for a large class of companions. We compare these posteriors to published orbital solutions for 25 stellar binaries from the Sixth Catalog of Orbits of Visual Binary Stars (or ORB6), recovering all companions at high significance and broadly consistent orbital separations. We then recover independent evidence supporting 94 of 120 tested Jovian exoplanetary systems from the NASA Exoplanet Archive, along with three known stellar companions, while ruling out one previously detected planet. We demonstrate that in cases like 14 Her b, the posteriors confirm the planetary nature of a signal using only Gaia and Hipparcos data. We find no false positives among 25 RV-quiet standard stars without significant Hipparcos-Gaia accelerations. Our method can break degeneracies inherent to proper-motion anomaly or excess-noise modeling alone by resolving orbital curvature within the Gaia baseline. The catalog and updated Octofitter are made publicly available to the community.

Gene regulatory mechanisms control cell differentiation and homeostasis but are often undetectable, particularly at the single-cell level. We introduce bayesReact, which quantifies regulatory activities from bulk or single-cell omics data. It is based on an unsupervised generative model, exploiting the fact that each regulator typically targets many genes sharing a sequence motif. Using mRNA expression data, we illustrate and evaluate bayesReact on microRNAs (miRNAs). It outperforms existing methods on sparse bulk data and improves activity inference on single-cell data. Inferred miRNA activities correlate with miRNA expression across pan-cancer TCGA and healthy GTEx tissue samples. The activities capture cancer-type-specific miRNA patterns, e.g., for miR-122-5p and miR-124-3p, which also correlate more strongly with their target genes than their measured expression. This includes a strong negative correlation between miR-124-3p and the anti-neuronal REST transcription factor in nervous system cancers. Analyzing single-cell data, bayesReact detects prominent miRNAs during murine stem cell differentiation, including miR-298-5p, miR-92-2-5p, and the Sfmbt2 cluster (miR-297-669). Furthermore, spatio-temporal inference shows increasing miR-124-3p activity in differentiating neurons during embryonic spinal cord development in mice. bayesReact enables large-scale hypothesis-generating screens for novel regulatory factors and the discovery of condition-specific activities. It is implemented as a user-friendly R package (https://github.com/JakobSkouPedersenLab/bayesReact).

The companion GL229B was recently resolved by Xuan et al. as a tight binary of two brown dwarfs (Ba and Bb) through VLTI-GRAVITY interferometry and Very Large Telescope-CRIRES+ radial velocity (RV) measurements. Here, we present Bayesian models of the interferometric and RV data in additional detail, along with an updated outer orbit of the brown dwarf pair about the primary. To create a model of the inner orbit with robust uncertainties, we apply kernel phases to the GRAVITY data, to address baseline redundancy in the raw closure phases. Using parallel tempering, we constrain the binary’s orbit using only VLTI-GRAVITY data, despite each epoch having low visibility-plane coverage and/or signal-to-noise ratio (SNR). We demonstrate very good agreement between the VLTI-GRAVITY and CRIRES+ data sets and find that the inner binary has a period of 12.1346 ± 0.0011 days, an eccentricity of 0.2317 ± 0.0025, and a total mass of 71.0 ± 0.4 M _jup, with Ba and Bb having masses of 37.7 ± 1.1 M _jup and 33.4 ± 1.0M _jup, respectively. With new Keck/NIRC2 astrometry, we update the outer orbit of GL229B around the primary. We find a semimajor axis of 42.9+3.0–2.4 au, an eccentricity of 0.736 ± 0.014, and a total mass for B of 71.7 ± 0.6 M _jup, consistent with that derived from the inner orbit. We find a mutual inclination of 31° ± 2 . ° 5, below the threshold for Kozai–Lidov oscillations. The agreement on the mass of Ba+Bb between the inner and outer orbits is an important test of our ability to model the RV, astrometry, and Hipparcos–Gaia proper-motion anomaly. Our methodological advances in handling interferometric data with low SNR and sparse UV coverage will benefit future observations of rapidly orbiting companions with VLTI-GRAVITY.

The companion GL229B was recently resolved by Xuan et al. (2024) as a tight binary of two brown dwarfs (Ba and Bb) through VLTI-GRAVITY interferometry and VLT-CRIRES+ RV measurements. Here, we present Bayesian models of the interferometric and RV data in additional detail, along with an updated outer orbit of the brown dwarf pair about the primary. To create a model of the inner orbit with robust uncertainties, we apply kernel phases to the GRAVITY data to address baseline redundancy in the raw closure phases. Using parallel tempering, we constrain the binary's orbit using only VLTI-GRAVITY data, despite each epoch having low visibility-plane coverage and/or SNR. We demonstrate very agreement the VLTI-GRAVITY and CRIRES+ datasets and find that the inner binary has a period of 12.1346$\pm$0.0011 days, eccentricity of 0.2317$\pm$0.0025, and total mass of 71.0$\pm$0.4 Mjup, with Ba and Bb having masses of 37.7$\pm$1.1Mjup and 33.4$\pm$1.0Mjup respectively. With new Keck/NIRC2 astrometry, we update the outer orbit GL229B around the primary. We find a semi-major axis of 42.9+3.0-2.4AU, eccentricity of 0.736$\pm$0.014, and a total mass for B of 71.7$\pm$0.6Mjup, consistent with that derived from the inner orbit. We find a mutual inclination of 31$\pm$2.5deg, below the threshold for Kozai-Lidov oscillations. The agreement on the mass of Ba+Bb between the inner and outer orbits is an important test of our ability to model RV, astrometry, and Hipparcos-Gaia proper motion anomaly. Our methodological advances in handling interferometric data with low SNR and sparse UV-coverage will benefit future observations of rapidly-orbiting companions with VLTI-GRAVITY.

Motivation: Regulatory constraints are crucial in maintaining tissue and cell integrity, and play important roles during developmental processes and environmental responses. Yet many regulatory mechanisms remain unobserved at the single-cell level and statistical inference may, in some cases, help elucidate their condition-specific activity and perturbation during disease progression. Results: We introduce bayesReact (BAYESian modeling of Regular Expression ACTivity), a generative model of motif occurrence across experimentally ranked sequences to infer motif-based regulatory activities. The method is evaluated for microRNAs (miRNAs), which perform post-transcriptional regulation through target mRNA destabilization and translational repression. Inferred miRNA activities positively correlate with the observed miRNA expressions in primary tumors from The Cancer Genome Atlas (TCGA) and mouse stem cells. The top miRNA activity profiles are as informative for TCGA cancer-type cluster identification as the top miRNA or mRNA expression profiles. The activity captures tissue-specific miRNA patterns observed in the matched expression, e.g., the expression of miR-122-5p in the liver and miR-124-3p in low-grade gliomas (LGG). We observe a negative association between the activity of the two miRNAs and their target gene expressions, including between the miR-124-3p activity and the anti-neuronal REST expression in LGG. bayesReact outperforms the existing method, miReact, on sparse count data, and shows a higher correlation with the miRNA expression in single-cell data. The method recovers temporal activities of prominent miRNAs during murine stem cell differentiation, including miR-298-5p, miR-92-2-5p, and the large Sfmbt2 cluster (miR-297-669). The bayesReact model is probabilistic and quantifies the uncertainty of all provided estimates. It is unsupervised and permits screens of bulk or single-cell data to identify condition-specific regulatory motif candidates. It further improves miRNA activity inference in single-cell data. Availability and implementation: bayesReact is implemented as an R-package, uses a Hamiltonian Monte-Carlo sampler for posterior approximation, and is available at https://github.com/astamr/bayesReact.

Parallel tempering (PT) is a class of Markov chain Monte Carlo algorithms that constructs a path of distributions annealing between a tractable reference and an intractable target, and then interchanges states along the path to improve mixing in the target. The performance of PT depends on how quickly a sample from the reference distribution makes its way to the target, which in turn depends on the particular path of annealing distributions. However, past work on PT has used only simple paths constructed from convex combinations of the reference and target log-densities. This paper begins by demonstrating that this path performs poorly in the setting where the reference and target are nearly mutually singular. To address this issue, we expand the framework of PT to general families of paths, formulate the choice of path as an optimization problem that admits tractable gradient estimates, and propose a flexible new family of spline interpolation paths for use in practice. Theoretical and empirical results both demonstrate that our proposed methodology breaks previously-established upper performance limits for traditional paths.

Bayesian inference for Continuous-Time Markov Chains (CTMCs) on countably infinite spaces is notoriously difficult because evaluating the likelihood exactly is intractable. One way to address this challenge is to first build a non-negative and unbiased estimate of the likelihood -- involving the matrix exponential of finite truncations of the true rate matrix -- and then to use the estimates in a pseudo-marginal inference method. In this work, we show that we can dramatically increase the efficiency of this approach by avoiding the computation of exact matrix exponentials. In particular, we develop a general methodology for constructing an unbiased, non-negative estimate of the likelihood using doubly-monotone matrix exponential approximations. We further develop a novel approximation in this family -- the skeletoid -- as well as theory regarding its approximation error and how that relates to the variance of the estimates used in pseudo-marginal inference. Experimental results show that our approach yields more efficient posterior inference for a wide variety of CTMCs.