Thank you Alba😘 Hope everything goes well with you in Boston

The manuscript is available online now at doi.org/10.1093/life... and we welcome discussions and hope this optimization process can help more research groups on the application of snRNAseq in the metabolism community! (6/n).

We built an open access tutorial (metabomicslab.github.io/snRNAseq-ana...) and have made the data and the preprocessing Snakemake pipeline accessible on Zenodo as well （https://zenodo.org/records/14725531). (5/n)

We applied the optimized workflow to analyze 67 samples from 4 major metabolic tissues in both human and mouse (201,411 nuclei). It showed enhanced cell-type resolution (e.g. adipocytes,) and the identification of biologically relevant pathways. (4/n)

After carefully benchmarking existing methods, we proposed an optimized workflow that includes:
✅ Removal of environmental RNA contamination (CellBender)
✅ Double cell detection (scDblFinder)
✅ Selection of integration methods (scIB framework) (3/n)

Single nucleus transcriptomics (snRNAseq) has become an alternative way to get single cell resolution of these tissues. However, it also brought analytic challenges given the high ambient RNA contamination and doublet rate. (2/n)

Metabolic homeostasis is regulated by multiple tissues, including adipose tissue, the liver, and the hypothalamus etc. However, traditional single cell methods cannot be directly applied to these tissues due to the fragile nature of cells, including adipocytes and neurons. (1/n)