Search# | Spectra | Database | PMTol | Decoy | EmpiricalFDR1 fixed | FactFDR2 fixed |
|---|
| | | | | |
| FactFDR(%) | p-value(%) 4 |
N
target
| EmpiricalFDR(%) |
|---|
I-1 | ISB-02 | ISB |
2.5 Da
| Rev | 2329/1009 | 5.8/4.4 | 10.9/30.9 | 2279/1024 | 3.9/5.7 |
I-2 | ISB-02 | ISB |
2.5 Da
| Shfl | 2339/1023 | 6.0/4.6 | 7.2/38.6 | 2279/1025 | 3.6/5.1 |
I-3 | ISB-02 | ISB+Yeast |
2.5 Da
| Rev | 1578/602 | 4.7/5.1 | 40.8/50.3 | 1583/596 | 5.1/3.9 |
I-4 | ISB-02 | ISB+Yeast |
2.5 Da
| Shfl | 1597/577 | 5.0/4.2 | 50.2/34.5 | 1589/588 | 4.8/5.6 |
I-5 | ISB-02+AB-TC | ISB+Yeast |
2.5 Da
| Rev | 1490/569 | 5.0/5.8 | 50.2/31.2 | 1480/553 | 4.5/4.3 |
I-6 | ISB-02+AB-TC | ISB+Yeast |
2.5 Da
| Shfl | 1488/530 | 5.0/4.0 | 50.2/28.7 | 1478/550 | 4.9/5.5 |
I-7 | ISB-02+AB-TC | ISB+AT |
2.5 Da
| Rev | 1320/441 | 4.6/4.1 | 36.6/38.0 | 1342/464 | 5.8/7.3 |
I-8 | ISB-02+AB-TC | ISB+AT |
2.5 Da
| Shfl | 1287/441 | 3.4/4.1 | 7.5/38.0 | 1342/464 | 6.8/7.3 |
Y-1 | Y-Small+AB-TC | Yeast+AT | 30 ppm | Rev | 2574/1988 | 1.0/1.3 | 50.1/22.7 | 2588/1759 | 1.0/0.5 |
Y-2 | Y-Small+AB-TC | Yeast+AT | 30 ppm | Shfl | 2554/1940 | 0.9/1.2 | 38.7/27.3 | 2620/1758 | 1.1/0.6 |
- All searches followed the standard TDA procedure except the step 2 for shuffled database searches. The results in columns labeled "FDR fixed" are obtained at empirical FDR threshold of 5% (the searches I-1 to I-8) or 1%(the searches Y-1 to Y-2). The results in columns labeled "FactFDR fixed" are obtained at factual FDR threshold of 5% (the searches I-1 to I-8) or 1%(the searches Y-1 to Y-2). The underlined characters represent either dummy spectra, dummy databases, or dummy tolerance. The first numbers in N
target
/FactFDR/FDR/p-value fields are from MS-GFDB, and the second from X!Tandem. Note that we do not aim to compare database search engines (i.e., MS-GFDB vs. X!Tandem). We only evaluate how FDR estimation via TDA is reliable and how the number of positive PSMs (or peptides) changes for different search strategies with different parameters or protocols.
- In contrast with popular belief, we did not observe a conservative estimation of FDR with shuffled decoy when compared to the reverse decoy database.
- 1the empirical FDR; 2the factual FDR; 3 the number of positive target PSMs; 4 Fisher p-value (see Table 2) - Fisher p-values less than 5% were emphasized with bold fonts.
