Maren Hackenberg

The paper introduces mmContext, a framework for multimodal contrastive learning of omics and text data, and demonstrates its utility in RNA-Seq and text integration tasks.

The paper introduces scCChain, a transformer-based framework for mapping spatial cell-cell communication programs in spatial transcriptomics data, identifying hotspots at single-cell and spot resolution.

The paper introduces a sparse dimensionality reduction workflow for analyzing single-cell cell-cell interactions using the Boosting Autoencoder approach, with a Jupyter notebook for implementation.

This paper proposes a deep learning method to align neural network architectures across species for single-cell data, revealing commonalities and differences in gene expression.

The paper fine-tunes biomedical language models to integrate text-based knowledge with scRNA-seq data, enhancing cell identity preservation and adding functional and disease-related context.

The paper evaluates discrepancies in dimensionality reduction techniques for time-series scRNA-seq data using synthetic data and finds that different methods may not reliably represent dynamics when used alone.

The paper introduces a boosting autoencoder method to incorporate structural assumptions into dimensionality reduction for single-cell RNA sequencing data, identifying small gene sets that explain latent dimensions.

This study evaluates deep generative models for integrating single-cell multi-omics data across varying sample sizes and provides recommendations for experimental design.

The paper develops an unsupervised deep learning model for two-photon imaging data that incorporates structural knowledge to improve interpretability and modeling of neural activity.