Fig. 4: Performance evaluation of high-performing method combinations on spike-in single-cell samples.
From: Benchmarking informatics workflows for data-independent acquisition single-cell proteomics

a Construction of the spike-in single-cell samples. b Selection of the high-performing method combinations based on the benchmarking results. c Performance of the selected high-performing method combinations on the spike-in single-cell samples. Metrics include: the adjusted Rand index (ARI) and partial area under curve (pAUC) values (indicated by dot sizes and colors), numbers of detected true negative (TN, blue bars), true positive (TP, green bars), false positive (FP, red bars), and false negative (FN, purple bars) proteins, as well as true positive rate (TPR, green lines) and false positive rate (FPR, red lines) values. Mappings of the serial numbers to detailed methods for each step are present in Fig. 2a. Average metrics of all the 1225 method combinations were used as a baseline (Base Avg). The data were processed starting from SR75. Differential analysis was performed between the C2Y1E and C1Y2E sample groups. Differential proteins are determined with p-value < 0.05 and |log2 FC | > log2 1.5.