| dc.description.abstract | Flow cytometry has become a powerful tool for studying complex biological processes at the level of individual cells. However, with the increasing number of markers measured in experiments, it has become challenging to analyze all possible combinations of marker expressions manually. To address these challenges, computational tools like clustering and dimensionality reduction have become essential for analyzing high-dimensional data in a more unbiased way. One issue that remains is the presence of batch effects, which are technical variations that can confound biological differences. These batch effects make it challenging to combine data from different experiments and utilize computational techniques. In this study, we implemented an algorithm called CytoNorm into an in-house developed data analysis pipeline to evaluate its effectiveness in normalizing batch effects in flow cytometry related datasets. Using both Uniform Manifold Approximation and Projection (UMAP) and expression pattern analysis of technical controls, we identified batch effects in two distinct datasets. Moreover, we showed that CytoNorm reduced the impact of batch effect on downstream clustering analysis of the cells. By addressing these challenges, our study aims to enhance the automation of flow cytometry analysis, allowing researchers to explore datasets in a more unbiased and unsupervised manner. By implementing these methods, the efficiency, reliability, and interpretation of flow cytometry data will be significantly enhanced, thereby advancing hypothesis testing in the field of human biology. | |
