Capturing Single-cell Heterogeneity via Data Fusion Improves Image-based Profiling, the Latest Result of Mutual Work of Sharif and Broad Institute Researchers

Dr. Mohammad H. Rohban, Faculty Member at Computer Engineering Department of Sharif University of Technology in a joint research project with Hamdah S. Abbasi, Shantanu Singh, and Anne E. Carpenter of Imaging Platform, Broad Institute of MIT and Harvard publish the result of their research on improving image-based profiling, “Capturing single-cell heterogeneity via data fusion improves image-based profiling”, in Nature Communication, an open access journal that publishes high-quality research from all areas of the natural sciences. Papers published by the journal represent important advances of significance to specialists within each field.

Single-cell resolution technologies warrant computational methods that capture cell heterogeneity while allowing efficient comparisons of populations. Bioinformatic readout technologies that have single-cell accuracy require the methods of digesting and indexing the population of cells that are capable of modeling the heterogeneity of single cells. A profile is a population of vector cells that provides comprehensive information on the biological status of a population of cells and allows comparison of cell populations through it. Comparison of cell populations has many applications in discovering drugs, identifying the function of genes, and discovering the mechanism of unknown drugs. In the new work of the researchers of SUT and Broad Institute, cell populations are summarized by adding features’ dispersion and covariances to population averages, in the context of image-based profiling. It is found that data fusion is critical for these metrics to improve results over the prior alternatives, providing at least ~20% better performance in predicting a compound’s mechanism of action (MoA) and a gene’s pathway.

Dr. Mohammad H. Rohban, Assistant Professor at Computer Engineering Department, conducts research on Image Analysis in Fluorescent Microscopy, High Throughput Biological Profiling, and Machine Learning at Interpretable Machine Learning Lab of Sharif University of Technology. To find out more you might visit his homepage at http://ce.sharif.edu/~rohban/

 

Reference:

Mohammad H. Rohban, Hamdah S. Abbasi, Shantanu Singh, & Anne E. Carpenter, Nature Communications volume 10, Article number: 2082 (2019)