Back

Spatial Transcriptomics Sequencing of Mouse Liver at 2 Micron Resolution Using a Novel Spatial DNA Chip

Ding, X.; Hoff, K.; Swaminathan, R.; Pollom, S.; Huang, T.; Li, X.; Zhou, G.; Bai, Z.; Yu, S.; Koutrouli, M.; Jensen, L. J.; Crnogorac, F.; Crnogorac, F.; Yu, S.; McGall, G.; Edwards, J.; Zhou, W.

2024-01-09 genomics
10.1101/2024.01.08.574734 bioRxiv
Show abstract

Spatial transcriptomics enables analysis of gene expression that is spatially resolved within a tissue section, making it possible to elucidate the relationship between individual cells within the context of the tissue. This transformative technology enables researchers to better understand gene function within the context of health tissue, developmental processes, and disease. In this study, we present an innovative spatial transcriptomics technology and data using a high-resolution DNA chip with a total capture region size of 6.5 x 6.5 mm containing 2 x 2 {micro}m features for spatial barcoding with no gaps between the features, thereby maximizing the capture area. These chips are manufactured at wafer scale using photolithography and are transferred to hydrogels, making them compatible with existing workflows for fresh frozen or paraffin-embedded samples. Herein, we examined a fresh frozen sample from an adult mouse liver. To analyze the data, we binned 10 x 10 features to represent a 20 {micro}m x 20 {micro}m capture area. We obtained 1.3 billion unique mapped reads, 68.78% sequencing saturation, with a median of 16,967 unique reads per region, indicating the potential for more unique reads with deeper sequencing. This high-resolution mapping of liver cell types and the visualization of gene expression patterns illustrate significant advancements in spatial sequencing technology.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.