Figure 1 | Scientific Reports

Figure 1

From: A machine learning approach toward automating spatial identification of LAG3+/CD3+ cells in ulcerative colitis

Figure 1

(a) Single plex assays for CD3 and LAG3 were compared to the dual plex assay to determine multiplex concordance with UC tissue shown in example (Top row, IF images; Bottom row, image analysis markup; Scale bars denote 50 μm). Image analysis for CD3+ cells (b,d) and LAG3+ cells (c,e) revealed strong concordance between assays. Image analysis was validated by comparing the algorithm to three trained and independent observers. This analysis revealed robust and repeatable image analysis algorithms for CD3+ (f), LAG3+ (g), and CD3+/LAG3+ dual positive (h) cells (a–e: N = 10 samples: 5 UC, 4 normal colon, 1 tonsil; f–h N = 10 samples all UC). Dashed lines denote 95% confidence intervals, colored by individual observer.

Back to article page