Table 2.

Performance metrics of FedProt and selected meta‑analyses using bacterial and human serum datasets

Dataset	Method	Mean difference	Maximal difference	FPa	FNa	Jaccard similarity coefficienta
Bacterial dataset	FedProt	4.45 × 10−13	3.56 × 10−12	0	0	1
	Fisher	3.86	25.24	0	1	0.998
	Stouffer	3.35	25.00	0	1	0.998
	REM	15.61	259.80	8	14	0.964
	RankProd	13.42	79.42	0	108	0.820
Human serum dataset	FedProt	1.36 × 10−13	6.59 × 10−13	0	0	1
	Fisher	0.50	2.79	2	6	0.922
	Stouffer	0.57	2.59	2	13	0.854
	REM	0.59	11.64	1	13	0.863
	RankProd	1.07	8.98	1	33	0.667

Mean and maximum absolute differences in negative log-transformed BH-adjusted P values, Jaccard similarity coefficients, and error rates (FPs and FNs) for the results of FedProt and selected meta-analysis approaches compared with centralized DEqMS results. The lowest differences are shown in bold font.

^aThresholds of |log₂FC| >0.5 and adjusted P value <0.05 (bacterial dataset) and |log₂FC| >0.25 and adjusted P value <0.05 (human serum dataset) were used for Jaccard similarity coefficients, FPs and FNs. Metrics use each method’s own |log₂FC| and BH-adjusted P values.