The latest version of Kasumi is now published in Nature Comms https://t.co/JRIqiYAtII Kasumi identifies patterns in tissue patches enabling analysis of disease progression and treatment response while providing insights into spatial coordination at cell-type or marker level

This work was led by @FCeccarelli96 in collaboration with @pl219_Cambridge, @SeanHoldenAI, @saezlab and Tanevski Lab.

We demonstrate TOAST on intra-, intersample, and temporal alignment in: 🧠 Human cortical layers (Visium) 🧫 Axolotl regeneration (Stereo-seq) 🐭 Locallization in mouse embryo development (seqFISH) 🎯 Various cancer types (IMC) … with state-of-the-art efficiency and performance.

TOAST quantifies spatial coherence using entropy in local neighborhoods—favoring alignments that keep the order of local spatial compositions. It also preserves neighborhood consistency—aligning spots with similar gene expression in the spatial neighborhood.

Alignment is an important step for data integration that can help gain insights into mechanisms, progression and structural changes in disease. When the spatial context is available it *has to* complement molecular similarity to yield more biologically plausible mappings.

🚨 New preprint: "Topography Aware Optimal Transport for Alignment of Spatial Omics Data" We present our new alignment framework TOAST https://t.co/EV1Jo0ThJQ

Together with @tanevski and @drbetge, we are looking for a computational postdoc interested in the combination of modeling tissue organization patterns from spatial proteomics data in a highly clinically relevant cancer context. Funding available:

📣 The much-revised manuscript of LIANA+, our all-in-one solution to study cell-cell communication from single-cell, spatial, and multi-omics technologies, is now published in @NatureCellBio https://t.co/KkHX3GVLn1

Tomorrow at BioQuant : Dissecting and overcoming different shades of cancer immune evasion by Benjamin Izar. Join us! Supported by @ResearchLifesci Hosted by: @DenisSchapiro , @JulioSaezRod and @tanevski Organised by MULTI-SPACE https://t.co/7CGxVulWuV

Our workflow, FLOP, exploring the impact of filtering, normalization, and differential expression on functional analysis is now published in @NAR_Open. https://t.co/zeohyVRcWz

I am particularly happy for this publication. It is a milestone in my recently started transition to a group leader. I am grateful to everyone involved - @Arezou_Rahimi for her amazing work and trust, @JulioSaezRod for his continuous support and the great collaboration with @GSK

In a very busy area with many great methods, we took a step forward. We generalize on localization and deconvolution and introduce new tasks, bridging information between samples and across modalities. We offer a flexible and scalable solution while achieving the best performance

🚨 We are looking for a PhD student 🚨 to join the Tanevski Lab and @DenisSchapiro Lab at the Heidelberg University Hospital, working at the intersection of machine learning and spatial omics as part of the Translational Spatial Profiling Center. Please RT

We welcome Julio Saez-Rodriguez @JulioSaezRod as our new Head of Research! Associate Director for EMBL-EBI Services Jo McEntyre @jomcentyre has been promoted to Deputy Director of EMBL-EBI & Rolf Apweiler is stepping back to Associate Director as he looks towards retirement.

Check it out here https://t.co/VJMiGGvAD3 (project number 71) or contact me directly. Please RT 🙂

If this sounds interesting to you and you happen to look for a new position, with @felixjhartmann we are hiring an interinstitutional postdoc to work with us in Heidelberg on capturing and exploring higher order tissue organization from highly multiplexed omics data.

If you like MISTy https://t.co/ySyNNPW9eZ, you'll really like Kasumi https://t.co/00A3UjAcmf. Kasumi is a multi-view explainable representation learning approach, capturing localized relationship-based patterns in tissues across conditions aiming towards facilitating translation

We have great scientific sessions lined up this year including "Artificial Intelligence / Machine Learning: Data Integration and Analytics for Biomolecular Imaging." We're thrilled to have speakers including Jovan Tanevski (Heidelberg University),

Amazing work by @DanielBDimitrov extending LIANA to now include also CCC approaches aimed at spatial data. I’m happy that i was able to help by showcasing MISTy. Huge thanks to @psl_schaefer for the adaptation, bringing MISTy to the Python leaning part of our community.

Integrating single-cell & spatial omics with different resolutions is key to understand tissue structure & function 🧬🔬 We had tackled this using multi-objective optimization - we just updated our paper and implemented an #rstats package https://t.co/JAU3FM0szn