The calculation of both ptmRS sequence probabilities and ptmRS site probabilities is based on the assumption that if the sequence of amino acids assigned to the respective MS/MS spectrum is correct and the MS/MS spectrum acquired did not result from fragmentation of two co-eluting peptides with similar m/z ratios, one of the putative isoforms must be the true assignment. Therefore, the sum of all sequence probabilities must equal 100 percent. When one or both of these assumptions is not correct, the calculated probabilities might not correlate with the true probabilities for correct modification localization.

The following figure shows the basic internal workflow of the ptmRS node for phosphorylation.

Internal ptmRS workflow example for phosphorylation
Internal ptmRS workflow example for phosphorylation

The ptmRS node performs the following steps:

Procedure

  1. On the basis of all potential modification sites in the peptide sequence, ptmRS calculates all putative phosphorylation isoforms.
  2. ptmRS divides the fragment spectrum into 100-Th windows and extracts the i most intense peaks for each window, where i is the peak depth used for the filtered fragment spectrum.
  3. For each phosphorylation isoform, S, and each peak depth, i, ptmRS calculates the list of theoretical fragment ions and matches them to the extracted peaks of the experimental spectrum.
  4. The probability, p, of matching a fragment peak purely by chance is calculated as follows:
  5. 𝑝=𝑁𝑃𝑒𝑎𝑘𝑠×𝑑𝑤
  6. where:
  7. NPeaks is the number of extracted peaks.
  8. d is the mass tolerance for matching peaks to the theoretical fragment ions.
  9. w is the extracted mass range.
  10. From the number of theoretical fragment ions, n, the number of matched fragment peaks, kS, and the probability, p, ptmRS calculates the probability, PS,i of matching kS or more peaks purely by chance as the cumulative binomial probability of matching kS or more peaks in n attempts1), 2):
  11. 𝑃𝑆,𝑖𝑋≥𝑘𝑠=𝑛𝑘𝑝𝑘×1−𝑝(𝑛−𝑘)𝑛𝑘=𝑘𝑠
  12. ptmRS reports the binomial peptide score of a specific isoform at peak depth, i, as follows:
  13. pRS𝑆,𝑖 =10×log10𝑃𝑆,𝑖 𝑋≥𝑘𝑆
  14. ptmRS determines the optimal peak depth ioptimal as the peak depth with the largest difference between the best-scoring phosphorylation isoform and the second-best-scoring isoform1).
  15. From the binomial peptide score at the optimal peak depth, ptmRS calculates an isoform confidence sequence probability and modification site probabilities. It calculates the ptmRS sequence probability as follows2):
  16. ptmRS sequence probability =𝑃𝑆,𝑖𝑜𝑝𝑡𝑖𝑚𝑎𝑙−1𝑃𝑆,𝑖𝑜𝑝𝑡𝑖𝑚𝑎𝑙−1𝑆
  17. The following table gives an example showing how ptmRS calculates sequence probabilities.
Calculating ptmRS sequence probabilities

Putative isoform

pRS score

1/P value

pRS sequence probability

AM(pS)PALGVM(pS)FSGVQ

AM(pS)PALGVMSF(pS)GVQ

AMSPALGVM(pS)F(pS)GVQ

121

118

59

1.26 × 1012

6.31 × 1012

7.94 × 1012

1.26 × 1012/1.89 × 1012=0.67

0.33

0.00

 

 

∑ =1.89 x 1012

∑ = 1.00

  1. The ptmRS node calculates the ptmRS site probability, for example, for a particular phosphorylation site by summing up the ptmRS sequence probabilities of those isoforms where the respective site is phosphorylated 3). The following table gives an example showing how ptmRS calculates site probabilities.
Calculating ptmRS site probabilities

pRS sequence probability

Putative isoform

0.67

0.33

0.00

AM

AM

AM

(pS)

(pS)

S

PALGVM

PALGVM

PALGVM

(pS)

S

(pS)

F

F

F

S

(pS)

(pS)

GVQ

GVQ

GVQ

ptmRS site probabilities

 

1.00

 

0.67

 

0.33