Skip to main content

Table 11 Comparison between peptide-level factual FDR and peptide-level FDR

From: False discovery rates in spectral identification

    

EmpiricalPepFDR1 fixed

  

FactPepFDR fixed

 

Search#

Spectra

Database

PMTol

N target

# peptides1

FactPepFDR(%)

p-value(%)

N target

# peptides

EmpiricalPepFDR(%)

I-5

ISB-02+AB-TC

ISB+Yeast

2.5 Da

1340/498

532/234

5.3/6.4

45.0/22.6

1335/479

529/226

4.9/4.4

I-22

ISB-All+AB-TC

ISB+Yeast

2.5 Da

10245/3688

758/304

6.6/6.6

7.0/4.3

9448/3596

696/292

2.4/4.1

I-14

ISB-02+AB-TC

ISB

2.5 Da

2088/907

727/333

6.7/5.1

10.4/50.3

1994/900

693/332

2.9/4.8

I-23

ISB-All+AB-TC

ISB

2.5 Da

16602/6676

1083/416

9.2/8.9

0.0/2.0

15663/6203

1015/385

1.7/1.3

Y-1

Y-Small+AB-TC

Yeast+AT

30 ppm

2574/1963

2355/1818

1.1/1.3

38.7/17.6

2556/1759

2339/1636

0.9/0.6

Y-8

Y-All+AB-TC

Yeast+AT

30 ppm

7867/5068

3142/2201

1.3/1.0

6.1/50.1

7121/5068

2849/2201

0.4/1.0

Y-5

Y-Small+AB-TC

Yeast

30 ppm

3209/2666

2916/2455

1.0/0.9

44.7/50.1

3209/2734

2916/2515

0.9/1.0

Y-9

Y-All+AB-TC

Yeast

30 ppm

10005/7309

3885/2987

1.0/0.9

50.0/44.9

10005/7309

3885/2987

0.9/0.9

  1. Score thresholds were determined using empirical/factual peptide-level FDR and used to calculate factual/empirical FDRs. For the searches I-5, I-22, I-14, and I-23, the peptide-level FDR thresholds were set to 5%, and for the remaining searches they were set to 1%. The search I-23 illustrates the difficulty of enforcing peptide-level FDR when searching small databases.
  2. 1Empirical peptide-level FDR.
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us