Additional file 6: of CellMapper: rapid and accurate inference of gene expression in difficult-to-isolate cell types
Year: 2016
Type: article
Abstract: Robustness of CellMapper to query gene choice; companion figure to Fig. 1b. To test the sensitivity of CellMapper to the choice of query gene, we repeated our analysis for kidney podocytes using ten distinct query genes (MAFB and the nine positive control genes in Additional file 13) and then assessed how well each analysis recovered an independent, experimentally-defined set of podocyte genes in mouse [23]. This plot shows the area under the precision-recall curve (AUPR) achieved when using each of the ten query choices. For comparison, we included a dotted line for the AUPR achieved by in silico nano-dissection when given all ten query genes at once (light blue dotted line; the area under the light gray line in Fig. 1b), or when given the original training set of 47 positive control and 97 negative control genes (dark blue dotted line; the area under the dark gray line in Fig. 1b). All ten single query gene searches for CellMapper resulted in a higher AUPR than in silico nano-dissection achieved when given all ten of these genes at once. MAFB was selected as the primary query gene for podocytes in this study (i.e. the red line in Fig. 1b) because it was used by the Genitourinary Developmental Molecular Anatomy Project [57] (GUDMAP) for all podocyte labeling and isolation in their large evaluation across kidney cell types. (PDF 20 kb)
Source: Figshare
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7138318908