Table 3.

Evaluation of different LFQ workflows based on the value of PCV of eight metaproteomic benchmark data sets. The performance of a workflow was categorized into four groups based on the following PCV values: superior (<0.14, double underline), good (0.14~0.3, single underline), fair (0.3~0.7, dotted underline) and poor (>0.7, no underline). The workflow abbreviations are provided in Table 2, and those benchmark data sets are in descending order by their total number of samples (the number of cases versus that of the controls, as listed in the brackets under each data set ID)

Workflow	PXD006224 (60:24)	PXD002882 (21:10)	PXD006129 (14:14)	PXD006070 (9:9)	PXD003028 (8:8)	PXD000987 (4:4)	PXD005929 (3:3)	PXD006810 (3:3)
BOX-MAD-SVD	0.05	0.15	0.47	0.38	0.44	0.64	0.27	1.31
BOX-EIG-KNN	0.26	0.48	0.32	0.19	0.26	0.22	0.10	0.28
BOX-QUA-BAK	0.43	0.76	0.65	0.54	0.54	0.46	0.22	0.64
BOX-VSN-CEN	0.09	0.02	0.14	0.09	0.10	0.09	0.04	0.43
CUB-EIG-KNN	0.29	0.52	0.36	0.21	0.29	0.25	0.11	0.31
CUB-MAD-SVD	0.05	0.15	0.47	0.38	0.44	0.64	0.27	1.31
CUB-RLR-CEN	0.75	0.93	0.85	0.59	0.61	0.53	0.24	0.55
CUB-VSN-BAK	0.11	0.18	0.15	0.76	0.12	0.11	0.05	0.60
LOG-EIG-CEN	0.10	0.24	0.13	0.12	0.12	0.09	0.03	0.12
LOG-PQN-ZER	0.90	0.95	1.27	1.07	0.87	0.68	0.29	1.34
LOG-TIC-BAK	0.21	0.31	0.23	0.45	0.21	0.17	0.09	0.72
LOG-VSN-SVD	0.06	0.15	0.48	0.39	0.45	0.65	0.27	1.31
NON-EIG-KNN	1.08	1.55	1.07	0.69	0.96	0.82	0.32	0.82
NON-MAD-SVD	0.05	0.15	0.47	0.38	0.44	0.64	0.27	1.30
NON-VSN-BAK	0.11	0.23	0.17	0.12	0.12	0.11	0.04	0.14
NON-ZSC-CEN	0.64	1.60	0.94	1.68	1.39	1.56	0.44	1.06
POW-CYC-BAK	1.25	2.95	2.46	1.20	4.55	10.83	1.00	6.08
POW-LOW-BAK	2.26	0.41	4.71	4.26	4.08	4.41	15.87	6.64
POW-TMM-ZER	1.68	1.07	1.11	1.47	4.07	2.65	11.16	1.07
POW-VSN-CEN	0.19	0.26	0.28	0.25	0.32	0.43	0.14	0.33