Paul Villoutreix
@paulvilloutreix.bsky.social
We develop machine learning frameworks to decode the geometry and topology of tissue development from spatial omics data | Junior Professor INSERM & Group Leader Turing Center for Living Systems | https://bioml.lis-lab.fr
Are you passionate about the mathematics of data science, computational geometry and topology and want to apply your skills to a beautiful biological question?
Come do a PhD with us!
You will help us understand of how boundaries are established and maintained in early heart development.
Come do a PhD with us!
You will help us understand of how boundaries are established and maintained in early heart development.
November 15, 2025 at 3:42 PM
Are you passionate about the mathematics of data science, computational geometry and topology and want to apply your skills to a beautiful biological question?
Come do a PhD with us!
You will help us understand of how boundaries are established and maintained in early heart development.
Come do a PhD with us!
You will help us understand of how boundaries are established and maintained in early heart development.
Had a very nice time visiting Granada for the annual heart ICDAR meeting! We presented Harshit Pateria's current work in collaboration with Robert Kelly following up from our recent publication on the second heart field www.nature.com/articles/s41...
October 17, 2025 at 3:51 PM
Had a very nice time visiting Granada for the annual heart ICDAR meeting! We presented Harshit Pateria's current work in collaboration with Robert Kelly following up from our recent publication on the second heart field www.nature.com/articles/s41...
Finally, principal components of jsPCA are directly interpretable in terms of spatially variable genes (SVG). We found that the top 3000 genes of the first principal component recovered 80% of the SVGs obtained by the widely used SPARK-X method.
September 19, 2025 at 6:50 AM
Finally, principal components of jsPCA are directly interpretable in terms of spatially variable genes (SVG). We found that the top 3000 genes of the first principal component recovered 80% of the SVGs obtained by the widely used SPARK-X method.
jsPCA can also learn a joint representations on some of the slices and use it to predict the domains in an unseen slice.
September 19, 2025 at 6:50 AM
jsPCA can also learn a joint representations on some of the slices and use it to predict the domains in an unseen slice.
The joint analysis of multiple slices by jsPCA generates common domains among datasets, in contrast to monoslice analysis..
September 19, 2025 at 6:50 AM
The joint analysis of multiple slices by jsPCA generates common domains among datasets, in contrast to monoslice analysis..
jsPCA achieves state-of-the-art accuracy for domain identification while reducing computation from hours to seconds, and scales to atlas-level datasets such as Stereo-seq.
September 19, 2025 at 6:50 AM
jsPCA achieves state-of-the-art accuracy for domain identification while reducing computation from hours to seconds, and scales to atlas-level datasets such as Stereo-seq.
Very happy to introduce jsPCA, a fast and interpretable computational framework for spatial transcriptomics that simultaneously identifies spatial domains and variable genes across multi-slice and multi-sample data.
September 19, 2025 at 6:50 AM
Very happy to introduce jsPCA, a fast and interpretable computational framework for spatial transcriptomics that simultaneously identifies spatial domains and variable genes across multi-slice and multi-sample data.
Finally, by performing targeted experiments, we showed that our approach was stable under severe class imbalance and (partially) missing class/modality pairs.
September 16, 2025 at 10:35 AM
Finally, by performing targeted experiments, we showed that our approach was stable under severe class imbalance and (partially) missing class/modality pairs.
We then applied our approach on the large public TCGA dataset and show that we could obtain high classification accuracy on up to 34 classes with three modalities
September 16, 2025 at 10:35 AM
We then applied our approach on the large public TCGA dataset and show that we could obtain high classification accuracy on up to 34 classes with three modalities
We first show on simulated data that our architecture is stable under class imbalance, minimal supervision and partially missing class/modality pairs.
September 16, 2025 at 10:35 AM
We first show on simulated data that our architecture is stable under class imbalance, minimal supervision and partially missing class/modality pairs.
To exploit labeled and unlabelled data, we proposed a semi-supervised approach, that enables latent space alignement.
September 16, 2025 at 10:35 AM
To exploit labeled and unlabelled data, we proposed a semi-supervised approach, that enables latent space alignement.
Very happy to share MODIS - a new Multi-Omics Data Integration method for Small and unpaired datasets!
We solve a long-standing problem in the field: how to perform data integration for small unpaired datasets
Our solution: learning with class imbalance on both reference and target datasets!
We solve a long-standing problem in the field: how to perform data integration for small unpaired datasets
Our solution: learning with class imbalance on both reference and target datasets!
September 16, 2025 at 10:35 AM
Very happy to share MODIS - a new Multi-Omics Data Integration method for Small and unpaired datasets!
We solve a long-standing problem in the field: how to perform data integration for small unpaired datasets
Our solution: learning with class imbalance on both reference and target datasets!
We solve a long-standing problem in the field: how to perform data integration for small unpaired datasets
Our solution: learning with class imbalance on both reference and target datasets!
What a nice week at CIRM in Marseille for the workshop "Mathematics of Single-Cell Data Analysis", organized by PEPR Digital Health and PEPR MathVives.
So many exciting mathematical approaches covering GRN inference, statistical testing and optimal transport.
Looking forward to the next edition!
So many exciting mathematical approaches covering GRN inference, statistical testing and optimal transport.
Looking forward to the next edition!
July 5, 2025 at 9:31 AM
What a nice week at CIRM in Marseille for the workshop "Mathematics of Single-Cell Data Analysis", organized by PEPR Digital Health and PEPR MathVives.
So many exciting mathematical approaches covering GRN inference, statistical testing and optimal transport.
Looking forward to the next edition!
So many exciting mathematical approaches covering GRN inference, statistical testing and optimal transport.
Looking forward to the next edition!
Very happy to see our paper on nested genes published! Very nice interdisciplinary collaboration with the Bertrand lab @ibdm.bsky.social - very combination of scRNASeq + smFISH to explore a fascinating question
=> www.sciencedirect.com/science/arti...
=> www.sciencedirect.com/science/arti...
July 1, 2025 at 4:26 PM
Very happy to see our paper on nested genes published! Very nice interdisciplinary collaboration with the Bertrand lab @ibdm.bsky.social - very combination of scRNASeq + smFISH to explore a fascinating question
=> www.sciencedirect.com/science/arti...
=> www.sciencedirect.com/science/arti...
Our latest work - MODIS: Multi-Omics Data Integration for Small and Unpaired Datasets - was in Singapore for the www.lmrl.org workshop at ICLR. Congrats Daniel Lepe-Soltero!!
Preprint: arxiv.org/abs/2503.18856
Preprint: arxiv.org/abs/2503.18856
April 28, 2025 at 6:16 AM
Our latest work - MODIS: Multi-Omics Data Integration for Small and Unpaired Datasets - was in Singapore for the www.lmrl.org workshop at ICLR. Congrats Daniel Lepe-Soltero!!
Preprint: arxiv.org/abs/2503.18856
Preprint: arxiv.org/abs/2503.18856
Finally we considered more complicated scenarios, where some of the data is missing in some of the class/modality pairs
March 25, 2025 at 6:58 AM
Finally we considered more complicated scenarios, where some of the data is missing in some of the class/modality pairs
We then evaluated the impact of class imbalance on multiple class imbalance scenarios (fully supervised and semi-supervised)
March 25, 2025 at 6:58 AM
We then evaluated the impact of class imbalance on multiple class imbalance scenarios (fully supervised and semi-supervised)
We first assessed the minimal amount of class supervision needed to align the latent representations of multiple modalities
March 25, 2025 at 6:58 AM
We first assessed the minimal amount of class supervision needed to align the latent representations of multiple modalities
Very happy to share our new preprint "MODIS: Multi-Omics Data Integration for Small and Unpaired Datasets"!
We proposed a new coupled variational auto-encoder architecture to solve the problem of data integration for small and unpaired datasets (diagonal integration)
arxiv.org/abs/2503.18856
We proposed a new coupled variational auto-encoder architecture to solve the problem of data integration for small and unpaired datasets (diagonal integration)
arxiv.org/abs/2503.18856
March 25, 2025 at 6:58 AM
Very happy to share our new preprint "MODIS: Multi-Omics Data Integration for Small and Unpaired Datasets"!
We proposed a new coupled variational auto-encoder architecture to solve the problem of data integration for small and unpaired datasets (diagonal integration)
arxiv.org/abs/2503.18856
We proposed a new coupled variational auto-encoder architecture to solve the problem of data integration for small and unpaired datasets (diagonal integration)
arxiv.org/abs/2503.18856
We have three posters, come see us at numbers #5 #11 #177!
December 12, 2024 at 5:48 PM
We have three posters, come see us at numbers #5 #11 #177!
Very exciting approaches for Spatial Biology at #ESSB2024 in Berlin
#SpatialOmics #SpatialBiology #Omics
#SpatialOmics #SpatialBiology #Omics
December 12, 2024 at 10:19 AM
Very exciting approaches for Spatial Biology at #ESSB2024 in Berlin
#SpatialOmics #SpatialBiology #Omics
#SpatialOmics #SpatialBiology #Omics
We were delighted to host @zardilis.bsky.social from Cambridge's Sainsbury Laboratory during the past two weeks! It was a fantastic opportunity to exchange ideas on the spatial biology of plants and beyond. Interesting new methods combining expression patterns, morphometric and physical features!!
November 25, 2024 at 10:28 AM
We were delighted to host @zardilis.bsky.social from Cambridge's Sainsbury Laboratory during the past two weeks! It was a fantastic opportunity to exchange ideas on the spatial biology of plants and beyond. Interesting new methods combining expression patterns, morphometric and physical features!!
We could use our approach to compare various conditions and genetic perturbations, revealing the role of T-box genes in the patterns of epithelial tension in the second heart field. 6/n
November 10, 2024 at 6:33 PM
We could use our approach to compare various conditions and genetic perturbations, revealing the role of T-box genes in the patterns of epithelial tension in the second heart field. 6/n
Performing clustering on the first dimension of the joint spatial PCA lead to highly relevant region in the tissue. These regions revealed two zones in the anterior and posterior part of the second heart field.. Coherent with the patterns of T-box gene expression. 5/n
November 10, 2024 at 6:33 PM
Performing clustering on the first dimension of the joint spatial PCA lead to highly relevant region in the tissue. These regions revealed two zones in the anterior and posterior part of the second heart field.. Coherent with the patterns of T-box gene expression. 5/n