Additional Settings

Chimeric search and reporting multiple PSMs

Sage is capable of searching for chimeric/co-fragmenting PSMs in any search configuration. This feature requires that report_psms > 1 to have any special effect. During the search phase, Sage will find the best candidate peptide for a spectrum (highest hyperscore), and then subtract the most intense MS2 peak matching each fragment ion from that peptide. Sage will then search the subtracted spectrum again, and repeat the above step until report_psms candidates have been found for the spectrum, or there are fewer than min_peaks MS2 peaks remaining in the spectra.

• chimera: Boolean. Search for chimeric/co-fragmenting PSMs (default: false).
• report_psms: Integer. The number of PSMs to report for each spectrum. Higher values might disrupt re-scoring, it is best to search with multiple values (default: 1).

Retention Time Prediction

By default, Sage will align retention times across all files in the search by matching RTs for the most confident PSMs shared between runs.

Global retention time alignment is performed using a modified algorithm based on Chen AT, et al. (opens in a new tab)

1. Transform all RTs into unit-less percentages (0.0 - 1.0)
2. Assume that the expected RT for a peptide can be estimated from the average RT across all runs
3. For each run, calculate a linear regression between the observed peptide RTs and the global average. Transform all PSM retention times by the regression parameters
4. Following global RT alignment, Sage will then build a RT prediction model, and integrate this information as a feature for PSM rescoring.

See Klammer et al. (opens in a new tab)

If, for some reason, you would prefer the above to be turned off, you can do via the predict_rt flag.

{
    // Oh no!
    "predict_rt": false
}