Abstract In women with high-grade serous ovarian cancer, chemotherapy remains the primary standard treatment, despite growing recognition of the disease as highly heterogeneous. Here, we examine the feasibility and clinical utility of comprehensive multimodal molecular profiling to inform treatment decisions. We analyze blood, single-cell and bulk tumor tissue, and malignant ascites using up to eleven technologies (DNA, RNA, protein, and functional assays) within a four-week turnaround time. Hypothetical treatment recommendations are altered for 76% of patients, and multi-omics-guided maintenance therapy is associated with prolonged overall survival in a subset of patients. Subsequent cohort analysis reveals distinct cellular and molecular profiles in ascites-derived single-cells compared to solid tumor tissue, unique per-patient ex vivo drug responses, and a marked increase in cancer cell heterogeneity following chemotherapy exposure. This coincides with genomic signature alterations in whole-genome-amplified patients. Our data suggest that molecularly guided treatments should be tested as adjuvant therapies prior to chemotherapy in the future.

Authors Francis Jacob, Rebekka Wegmann, Joanna Ficek-Pascual, Ulrike Lischetti, Jack Kuipers, Stéphane Chevrier, Michael Prummer, Nora C. Toussaint, Ilaria Alborelli, Ricardo Coelho, Ruben Casanova, Sandra Goetze, Gabriele Gut, Monica-Andreea Baciu-Drăgan, Anne Bertolini, Ximena Bonilla, Jon Brugger, Magdalena M. Brune, Byron Calgua, Stefanie Engler, Cinzia Esposito, Pedro F. Ferreira, Andrea Jacobs, Flavio C. Lombardo, Paola Malsot, Julien Mena, Nicola Miglino, Emanuela S. Milani, Jacobo Sarabia Del Castillo, Sujana Sivapatham, Bettina A. Sobottka-Brillout, Tanmay Tanna, Tibor A. Zwimpfer, The TumorProfiler Consortium, Rudolf Aebersold, Marina Bacac, Niko Beerenwinkel, Christian Beisel, Bernd Bodenmiller, Viktor H. Koelzer, Kjong-Van Lehmann, Mitchell P. Levesque, Holger Moch, Simone Muenst, Lucas Pelkmans, Markus G. Manz, Gunnar Rätsch, Franziska Singer, Berend Snijder, Alexandre P. A. Theocharides, Markus Tolnay, Andreas Wicki, Bernd Wollscheid & Viola Heinzelmann-Schwarz

Submitted Nature Communications

Link DOI

Abstract 10x Genomics VisiumHD enables spatial transcriptomics at 2 µm x 2 µm resolution but exhibits slide-specific, non-periodic striping artifacts due to lane-width variability. These multiplicative row/column effects distort bin total counts and can bias downstream analyses. The state-of-the-art destriping approach is the normalization procedure used as a preprocessing step in bin2cell; it applies sequential high-quantile row- then column-wise normalization, which is asymmetric and can introduce edge effects/macro-stripes and distortions of large-scale total-count structure. We propose a statistical destriping approach that leverages nuclei segmentation from the co-registered H&E image. Assuming transcript abundance is constant within each nucleus, we model bin counts with a negative binomial distribution whose mean is a product of a nucleus-specific concentration and row- and column-specific stripe-factors reflecting lane-width variation. We fit all parameters in a generalized linear modeling framework with cross-validated regularization on stripe-factors and iterative dispersion estimation, and use the fitted parameters to correct the observed counts into a destriped image. On synthetic data with known ground truth, our method improves stripe-factor estimation accuracy and reduces error in corrected counts relative to bin2cell and bin2cell-derived baselines. Across four public VisiumHD slides, it consistently lowers striping intensity while substantially better preserving biological signal present in the large-scale global count structure and avoiding the artifacts introduced by other methods.

Authors Paola Malsot, Malte Londschien, Valentina Boeva, Gunnar Rätsch

Submitted Bioinformatics

Link DOI

Abstract Knowing which features of a multivariate time series to measure and when is a key task in medicine, wearables, and robotics. Better acquisition policies can reduce costs while maintaining or even improving the performance of downstream predictors. Inspired by the maximization of conditional mutual information, we propose an approach to train acquirers end-to-end using only the downstream loss. We show that our method outperforms random acquisition policy, matches a model with an unrestrained budget, but does not yet overtake a static acquisition strategy. We highlight the assumptions and outline avenues for future work.

Authors Fedor Sergeev, Paola Malsot, Gunnar Rätsch, Vincent Fortuin

Submitted SPIGM ICML Workshop

Link DOI