Image: General conceptual framework and validation of gPRINT (IMAGE)
Caption
(A) Selected human single-cell transcriptome profiles from HCL and other public datasets were utilized to train and validate gPRINT. The human single-cell data encompasses 159,302 cells from 26 tissues and 5 platforms. (B) For each cell, a neural network of that cell and its “gene print” was constructed for supervised learning of gPRINT using known cell labels from each tissue’s transcriptome atlas. (C) The performance of gPRINT was tested using both an internal human dataset and an external test dataset, which contains single-cell transcriptome data from multiple tissues. For the external human test dataset, gPRINT was validated at the levels of cell type, hybrid hierarchy type, and cell subtype
Credit
X Yan R, Fan C, Gu S, Wang T, Yin Z, Chen X
Usage Restrictions
none
License
Original content