MerQuaCo: A Tool for Quality Control in Spatial Transcriptomics
Author Information
Author(s): Martin Naomi, Olsen Paul, Quon Jacob, Campos Jazmin, Cuevas Nasmil Valera, Nagra Josh, VanNess Marshall, Maltzer Zoe, Gelfand Emily C, Oyama Alana, Gary Amanda, Wang Yimin, Alaya Angela, Ruiz Augustin, Reynoldson Cade, Bielstein Cameron, Pom Christina Alice, Huang Cindy, Slaughterbeck Cliff, Liang Elizabeth, Alexander Jason, Ariza Jeanelle, Malone Jocelin, Melchor Jose, Colbert Kaity, Brouner Krissy, Shulga Lyudmila, Reding Melissa, Latimer Patrick, Sanchez Raymond, Barta Stuard, Egdorf Tom, Madigan Zachary, Pagan Chelsea M, Close Jennie L, Long Brian, Kunst Michael, Lein Ed S, Zeng Hongkui, McMillen Delissa, Waters Jack
Primary Institution: Allen Institute for Brain Science
Hypothesis
How can we effectively detect and quantify imperfections in spatial transcriptomics datasets?
Conclusion
The MerQuaCo tool facilitates rigorous assessment of data quality in spatial transcriptomics, improving the accuracy of cell type identification.
Supporting Evidence
- MerQuaCo detects and quantifies imperfections in spatial transcriptomics datasets.
- Common imperfections include tissue loss and varying detection probabilities.
- The tool streamlines the analysis of large spatial datasets.
- Data loss was observed in 31% of sections analyzed.
- Detection efficiency varied significantly across the field of view.
- Manual inspection revealed additional imperfections not detected by the tool.
- Transcript density varied significantly across sections from the same mouse.
- MerQuaCo can be adapted for use with other spatial transcriptomics platforms.
Takeaway
This study created a computer program that helps scientists check the quality of images used to study genes in brain tissue, making it easier to find mistakes.
Methodology
The study involved collecting data from 641 mouse brain sections and developing a pixel classifier to identify common imperfections in the data.
Potential Biases
Potential biases may arise from subjective manual inspections that complement the automated tool.
Limitations
The tool may not detect imperfections that are only evident when comparing multiple sections.
Participant Demographics
Adult mouse brain sections were used in the study.
Statistical Information
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website