Abstract
Motivation Spatially resolved omics technologies are enhancing our understanding of tissues architecture. Despite major technological improvements, gaining in spatial resolution becomes experimentally expensive, while generating spatial landscapes at moderate resolution combined with computational methods for depixelating data represent a cost-effective strategy allowing to enlarge the number of experiments to be performed.
Results We have developed a computational strategy able to gain several-folds of resolution by inferring pseudo-interstitial pixels from their closest neighbors. This strategy has been validated in the context of public spatial transcriptomics data issued from melanoma, and human brain cortex tissue sections, by improving the identification of distinct tissue substructures. Furthermore, this methodology has been used for enhancing the resolution of consecutive sections collected from human brain organoids, as a way to demonstrate that a moderate resolution technology, combined with spatial depixelation processing allows to properly discern molecular tissue structures even in small tissues.
Contact mmendoza{at}genoscope.cns.fr
Competing Interest Statement
The authors have declared no competing interest.