Skip to main content

Table 5 Comparison between two FDR formulas

From: False discovery rates in spectral identification

Search#

Spectra

Database

PMTol

Formula

EmpiricalFDR fixed

FactFDR fixed

     

N target

FactFDR(%)

p-value(%)

N target

EmpiricalFDR(%)

I-3

ISB-02

ISB+Yeast

2.5 Da

1

1578/602

4.7/5.1

40.8/50.3

1583/596

5.1/3.9

I-12

ISB-02

ISB+Yeast

2.5 Da

2

1452/550

2.3/2.5

0.0/2.9

1583/596

9.6/7.4

I-5

ISB-02+AB-TC

ISB+Yeast

2.5 Da

1

1490/569

5.0/5.8

50.2/31.2

1480/553

4.5/4.3

I-13

ISB-02+AB-TC

ISB+Yeast

2.5 Da

2

1387/502

2.8/3.2

0.6/15.7

1480/553

8.7/8.3

Y-1

Y-Small+AB-TC

Yeast+AT

30 ppm

1

2574/1988

1.0/1.3

50.1/22.7

2588/1759

1.0/0.5

Y-4

Y-Small+AB-TC

Yeast+AT

30 ppm

2

2453/1626

0.8/0.8

27.8/36.2

2588/1759

2.0/1.0

  1. The Formula field in the fifth column specifies the formula for the FDR calculation: 1 for N decoy /N target and 2 for 2 · N decoy /(N target +N decoy ). For all searches, the standard TDA procedure was followed except the step 4 for searches using formula 2. The searches I-12 and I-13 show that using formula 2 results in conservative FDR estimation.