The algorithm

The cumulative binomial probability (P) (Equation 1) represents the probability of randomly matching at least the given number of fragment ions to the tandem mass spectrum, which is calculated by using the total number of fragment ions for the given peptide (N), the number of ions matched to the spectrum (n) and the probability of matching a peak (p).

To make the scores more human readable, the result from Equation 1 (Pi) is transformed to a logarithmic scale and the final PeptideScore(i) is obtained (Equation 2):

This entire process is repeated in 10 iterations. In each loop different peak depths (i) are chosen to calculate the cumulative binomial probability and finally ten different are generated. A different weight is assigned for each score (1 = 0.5; 2 = 0.75; 3 = 1; 4 = 1; 5 = 1; 6 = 1; 7 = 0.75; 8 = 0.5; 9 = 0.25; and 10 = 0.25), and then a weighted average score called peptide score is generated.

The pipeline:

###Fragmentation annotation rules

For the matched peaks, a rule-based expert system is used to filter them. These are the main rules used:

• Ignore annotations for internal fragments, immonium ions and side chain ions
• If fragmentation is not coming from HCD or CID instruments, remove the parent ions from the candidate list
• If the user chooses water loss or other neutral loss, add annotation with water loss or neutral loss into the candidate list
• If peptide charge is 1+, add 1+ annotation charge into candidate list
• If peptide charge is 2+, add 1+ and 2+ annotation charge into candidate list
• If peptide charge is greater than 2+, Ignore 3+ annotation charge, only add 1+ and 2+ annotation charge into the candidate list
• Align different prior weights for different annotation ion types: b ions (100), y ions (99), parent ions (98), a ions (97), x ions (96), c ions (95), z ions (94). Choose the highest weighted annotation, if there is more than one candidate
• The m/z chart shows the delta mass between the annotated ion and the experimental mass of the peak