First, check out my recent bioRxiv preprint w/ @oweinerlab.bsky.social: www.biorxiv.org/content/10.1...
We find contractility shifts the proportion of implantation-competent embryos from young and aged females. Keep reading this for more info!

Since I started my PhD in 2013, I have been tantalised by a question: can we measure or evaluate the transcriptional and mechanical states of cells in a tissue at the same time? I pursued this question as a postdoc and then as a PI! A🧵on the answer we have found 👇:

9. While we focused on scRNA-seq, early results suggest CONCORD’s applicability to spatial transcriptomics & scATAC-seq. It’s open-source in Python. See galleries: qinzhu.github.io/Concord_docu.... Try it: github.com/Gartner-Lab/....

8. CONCORD integrates datasets based solely on gene co-expression structures—without assuming batch-effect models or shared cellular states. It achieves robust alignment even with minimal dataset overlap, offering faster and scalable analysis from small studies to atlas-scale projects.

7. The key innovation leading to this performance gain is a probabilistic, dataset- and neighborhood-aware sampling strategy tailored for self-supervised contrastive learning, enabling the generation of coherent, denoised and high-resolution latent cell representations across diverse datasets:

6. CONCORD also provides biologically interpretable encodings representative of context-dependent gene-expression programs.

5. An intestinal development atlas demonstrated CONCORD's ability to resolve complex topologies, including differentiation trajectories intertwined with cell cycle loops (click here for interactive 3D UMAP! qinzhu.github.io/Concord_docu...):

4. We validated CONCORD on an embryonic atlas of C. elegans and C. briggsae from @jisaacmurray. CONCORD captured known bifurcations, resolved subtle cell states, and uncovered lineage convergence events missed by other methods (click here for interactive 3D UMAP! qinzhu.github.io/Concord_docu...):

3. We evaluated CONCORD on datasets with topologies such as continuous trajectories, loops, clusters and hierarchical trees. Using scIB metrics, geometric measures, and TDA, we assessed its ability to preserve biological structure. CONCORD outperforms existing methods:

2. CONCORD is a simple but powerful ML framework tackling data integration, dimensionality reduction, and denoising in single-cell analysis, developed by @qinzhu1. Read our bioRxiv preprint: doi.org/10.1101/2025...