Just as it is difficult to understand a conversation without knowing its context, it may be difficult for biologists to understand the importance of gene expression without knowing the cell environment. To solve this problem, researchers at Princeton Engineering have developed a method of elucidating the cell environment so that biologists can better understand information about gene expression.
Researchers, led by computer science professor Ben Raphael, hope the new system will open the door to detecting rare cell types and selecting cancer treatment options with new precision. Raphael is the senior author of an article describing the method, published May 16 Natural science methods.
The basic technique of binding gene expression with the cell environment, called spatial transcriptomics (ST) has existed for several years. Scientists are breaking down tissue samples on a microscale grid and link each site on the grid with information about gene expression. The problem is that modern computing tools can only analyze spatial models of gene expression in two dimensions. Experiments involving multiple pieces of a single tissue sample, such as an area of the brain, heart, or tumor, are difficult to synthesize into a complete picture of cell types in tissue.
Princeton’s method, called PASTE (for the probabilistic alignment of ST experiments), combines information from multiple slices taken from a single tissue sample, providing a three-dimensional view of gene expression in a tumor or developing organ. If the coverage of the sequence in the experiment is limited due to technical or cost, PASTE may also combine information from multiple tissue sections into one two-dimensional consensus section with richer information about gene expression.
“Our method was motivated by the observation that biologists often conduct multiple experiments with the same tissue,” Raphael said. “Now these repeated experiments are not exactly the same cellsbut they are of the same fabric and therefore must be very similar. ‘
The technique team can align multiple slices from a single tissue sample, classifying cells based on their gene expression profiles while maintaining the physical location of the cells in the tissue.
The project began in the summer of 2020 after Max Land, a math concentrator from Princeton Class 2021, took Raphael’s course “Algorithms in Computational Biology”. Excited by the rapid evolution and opportunity to improve understanding of human health and disease, Land asked Raphael to participate in research and began working on code to develop what became the PASTE method. He was advised by Raphael and lead author Ron Zeira, a former Princeton doctoral student who is now a researcher at Verily, a company that specializes in precision health.
The work was the focus of Land’s senior dissertation, and he co-authored the paper with Zeira, Raphael, and Alexander Strzelkowski, Ph.D. student. Now a computing biologist with the Sloan Catering Memorial Cancer Center in New York City, Land said Zeira and Raphael’s mentorship has played an important role in his pursuit of a research career.
The team developed their method using simulation gene expression data from a spatial transcriptomics study of a breast tumor where a correspondence between tissue sections had previously been established. They then evaluated the method based on data collected from brain samples prefrontal cortexwhich has a known structure consisting of layers of different cell types with unique gene expression signatures.
The researchers also applied PASTE to data collected from skin cancer biopsies in four different patients. Preliminary analysis of these data suggested that complex patchwork cells with a high degree of mixing of cancer and healthy cells. The PASTE method, however, showed that the apparent low spatial coherence in the three patient samples was probably due to the low sequence coverage in the experiments. A new analysis showed that the cells were grouped into more contiguous clusters, a more biologically plausible scenario.
“Once we integrate several of these slices and effectively increase data coverage, we get more spatially coherent groups of cells, which is smarter than any cell type randomly located in tissues,” Zeira said.
So far, the largest data set analyzed by the team was a sample of heart tissue with nine slices, but they focused on experiments with mouse embryos that included more than 30 pieces. Apart from computational considerations, experiments on spatial transcriptomics on such a scale remain expensive for many laboratories, Raphael said.
However, he added, “we hope that having a tool like PASTE will encourage more researchers to repeat the experiment, because now they can actually use information from additional slices as they previously could not.”
Ron Zeira et al., Alignment and integration of spatial transcriptomics data, Natural science methods (2022). DOI: 10.1038 / s41592-022-01459-6
Citation: A new method combines data to create a three-dimensional map of cell activity (2022, May 16), obtained May 16, 2022 from https://phys.org/news/2022-05-method-melds-d-cells .html
This document is subject to copyright. Except for any honest transaction for the purpose of private study or research, no part may be reproduced without written permission. The content is provided for informational purposes only.