SORT disentangles multilevel gene programs across heterogeneous spatial transcriptomic atlases
Understanding how transcriptional programs are organized across biological contexts is essential for deciphering tissue development, disease progression, and species-specific adaptations. Spatial transcriptomic atlases spanning patients, developmental stages, and species provide an opportunity to resolve gene expression programs (GEPs) that are sample-specific, group-shared, or region-associated,…
The SORT framework disentangles multilevel gene programs across diverse spatial transcriptomic atlases, facilitating the comprehension of tissue development, disease progression, and species-specific adaptations. By jointly recovering program-specific and shared gene expression patterns, SORT employs matrix factorization to identify distinct and overlapping activities, orthogonality to diminish program redundancy, and graph total variation to maintain spatial boundaries.
In simulations involving significant patient variation, SORT demonstrated greater accuracy in recovering patient-specific, subtype-shared, and region-associated programs compared to existing techniques.
SORT analysis of 108 pancreatic ductal adenocarcinoma sections revealed tumor and stromal programs with unique and intersecting activities, which were subsequently linked to survival outcomes across four distinct external cohorts. In a cross-species cerebellar atlas, SORT successfully distinguished conserved layer-associated programs from species-specific activity variation.
Additionally, across eight stages of mouse organogenesis, SORT unveiled gene expression programs (GEPs) characterized by anatomical redeployment and graded spatial activity, as well as stage-dependent local co-deployment of distinct programs.
Written by urgent.news from bioRxiv's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
