A bstract: Probabilistic models have provided the underpinnings for state-of-the-art performance in many single-cell omics data analysis tasks, including dimensionality reduction, clustering, differential expression, annotation, removal of unwanted variation, and integration across modalities. Many of the models being deployed are amenable to scalable stochastic inference techniques, and accordingly they are able to process single-cell datasets of realistic and growing sizes. However, the community-wide adoption of probabilistic approaches is hindered by a fractured software ecosystem resulting in an array of packages with distinct, and often complex interfaces. To address this issue, we developed scvi-tools ( https://scvi-tools.org ), a Python package that implements a variety of leading probabilistic methods. These methods, which cover many fundamental analysis tasks, are accessible through a standardized, easy-to-use interface with direct links to Scanpy, Seurat, and Bioconductor workflows. By standardizing the implementations, we were able to develop and reuse novel functionalities across different models, such as support for complex study designs through nonlinear removal of unwanted variation due to multiple covariates and reference-query integration via scArches. The extensible software building blocks that underlie scvi-tools also enable a developer environment in which new probabilistic models for single cell omics can be efficiently developed, benchmarked, and deployed. We demonstrate this through a code-efficient reimplementation of Stereoscope for deconvolution of spatial transcriptomics profiles. By catering to both the end user and developer audiences, we expect scvi-tools to become an essential software dependency and serve to formulate a community standard for probabilistic modeling of single cell omics.

Molecular generative artificial intelligence is drawing significant attention in the drug design community, with several experimentally validated proof of concepts already published. Nevertheless, generative models are known for sometimes generating unrealistic, unstable, unsynthesizable, or uninteresting structures. This calls for methods to constrain those algorithms to generate structures in drug-like portions of the chemical space. While the concept of applicability domains for predictive models is well studied, its counterpart for generative models is not yet well-defined. In this work, we empirically examine various possibilities and propose applicability domains suited for generative models. Using both public and internal data sets, we use generative methods to generate novel structures that are predicted to be actives by a corresponding quantitative structure–activity relationships model while constraining the generative model to stay within a given applicability domain. Our work looks at several applicability domain definitions, combining various criteria, such as structural similarity to the training set, similarity of physicochemical properties, unwanted substructures, and quantitative estimate of drug-likeness. We assess the structures generated from both qualitative and quantitative points of view and find that the applicability domain definitions have a strong influence on the drug-likeness of generated molecules. An extensive analysis of our results allows us to identify applicability domain definitions that are best suited for generating drug-like molecules with generative models. We anticipate that this work will help foster the adoption of generative models in an industrial context.

Despite growing interest and success in automated in-silico molecular design, questions remain regarding the ability of goal-directed generation algorithms to perform unbiased exploration of novel chemical spaces. A specific phenomenon has recently been highlighted: goal-directed generation guided with machine learning models produce molecules with high scores according to the optimization model, but low scores according to control models, even when trained on the same data distribution and the same target. In this work, we show that this worrisome behavior is actually due to issues with the predictive models and not the goal-directed generation algorithms. We show that with appropriate predictive models, this issue can be resolved, and molecules generated have high scores according to both the optimization and the control models.

A bstract: During the COVID-19 pandemic, epidemiological modelling has played a key role in public debate and policy making for anticipating the epidemic trajectory, as well as proposing and evaluating non-pharmaceuticals interventions. Despite its importance, evaluations of models’ ability to accurately represent the evolution of the disease remain scarce. Robust and systematic evaluation is needed to assess models. We investigate the following research question : were the COVID-19 scenarios proposed by modellers during the pandemic to policy-makers relevant for decision making ? To answer this, we conduct a retrospective assessment of modelling reports which guided policy response in France in 2020-2022. After systematically verifying the scenarios hypotheses (e.g., exclusion of no-lockdown scenarios when a lockdown was effectively in place), we find that out of 10 reports, reality was below the best-case scenario in 6 reports; within the best-case / worst case scenarios range in 3 reports; above the worst-case scenario in 1 report. Best-case scenarios were the closest to reality, but often came from report with a large span between best-case and worst-case scenarios beyond 2 weeks, precluding certainty about future outcomes at the time of publishing. Our results hint a systematic overestimation bias for these particular models used to anticipate epidemic evolution, which can be of importance if such models are used to contractually estimate the effectiveness of non pharmaceutical interventions. To our knowledge, this is the only national systematic retrospective assessment of COVID-19 pandemic scenarios assessing hospital burden; such an approach should be reproduced in other countries whenever possible. Graphical Reality (black line) compared to prospective scenarios (colored lines) which informed policy during the COVID-19 pandemic in France for Intensive Care Units (top) and New Hospital Admissions (bottom). Colors indicate the error between reality and scenarios, expressed as a percentage of the 1st wave peak (horizontal dashed line).