Skip to content

Peak picking & fitting

Peak fitting and deconvolution

Fit overlapping peaks to a lineshape model to recover intensities you can trust in a downstream quantitative fit.

Peak height is a fine measure of intensity right up until peaks overlap, at which point it silently becomes a measure of your neighbour. In a crowded amide region at the intensities that matter, that error is systematic, not random, and it survives into every rate you fit afterwards.

Lineshape fitting solves the overlap explicitly: peaks are fitted as a cluster to a Lorentzian, Gaussian or Voigt model, so intensity is apportioned between neighbours according to the model rather than assigned to whichever maximum happened to be highest. For a relaxation or CPMG series, the same cluster is fitted across every plane with shared position parameters, which is what makes the resulting decay curve meaningful.

Fitted intensities come back with uncertainties, and those uncertainties propagate. A rate fitted from fitted intensities carries an error bar that reflects the lineshape fit, not just the scatter of the points.

What it needs

  • Spectrum or pseudo-3D series
  • Peak list defining what to fit

What you get

  • Fitted intensities and volumes with uncertainties
  • Per-peak fit diagnostics

What does the work

  • PICA

    Built in-house

    Cluster deconvolution fitting

  • peakipy

    Open source, redistributed under its licenceMIT

    Lorentzian/Gaussian deconvolution

  • FuDA

    Open source, redistributed under its licenceMIT-equivalent: written consent from D. Flemming Hansen

    Cluster lineshape fitting

Related in Peak picking & fitting