apt-peak-background

name	apt-peak-background
description	Background-corrected peak counting — fit a linear background under a single mass-spectrum peak from interactively brushed bins and get its corrected counts/percentage. Use when the user wants accurate peak counts with background subtraction, signal-to-background, or to build a table of corrected peaks.

Compute background-corrected counts for individual mass-spectrum peaks by fitting a linear background from manually selected background bins. Use the MATLAB MCP server.

Prerequisites

The MATLAB workspace must contain:

• pos — loaded atom probe data in RAW (undecomposed) format (from /apt-load)
• spec — an open, ranged mass spectrum (from /apt-spectrum or /apt-massspec-auto)
• rangeTable — extracted from the spectrum. If missing, create it:

  rangeTable = rangesExtractFromMassSpec(spec);
  

If pos is decomposed, tell the user this workflow needs the raw pos file. If no ranged spectrum exists, run /apt-spectrum or /apt-massspec-auto first.

Steps

1. Background-correct one peak (interactive)

The mass spectrum figure must be the current figure and zoomed to the peak of interest. Because the function requires interactive point selection, the user must run this line in the MATLAB command window (it will not work fired blindly from a script):

[peakData] = peakBGCorrectedCount(pos, rangeTable, ionName);

ionName must be the specific-isotope form, e.g. '14N 12C2 +'. Tell the user to select 4 points in the spectrum — two before and two after the peak — marking the bins used for the linear background fit. The function plots the peak with its fitted background and returns the peakData struct (fields include counts, pct, loc).

Optional arguments (confirmed signature peakBGCorrectedCount(pos, rangeTable, ionName, boundary, options)):

• boundary — Da before/after the peak for the background range, scalar or [lo hi] (default 0.5)
• a figure flag (true/false) to suppress the plot
• options 'r' (add a fictional range and report signal-to-background) or 'a' (auto-create that range by balancing missed atoms vs. background atoms)

2. Start a results table (first peak)

rangeTable = rangesExtractFromMassSpec(spec);
peakName        = rangeTable.rangeName(rangeTable.mcbegin < peakData.loc & rangeTable.mcend > peakData.loc);
peakChargeState = rangeTable.chargeState(rangeTable.mcbegin < peakData.loc & rangeTable.mcend > peakData.loc);
peakConc = table(peakName, peakChargeState, peakData.loc, peakData.counts, peakData.pct, ...
    'VariableNames', {'peakName', 'peakChargeState', 'peakLoc', 'peakCounts', 'peakPct'});
clearvars peakName peakChargeState

3. Append each additional peak

Re-run step 1 on the next peak, then append:

rangeTable = rangesExtractFromMassSpec(spec);
peakName        = rangeTable.rangeName(rangeTable.mcbegin < peakData.loc & rangeTable.mcend > peakData.loc);
peakChargeState = rangeTable.chargeState(rangeTable.mcbegin < peakData.loc & rangeTable.mcend > peakData.loc);
peakNew = table(peakName, peakChargeState, peakData.loc, peakData.counts, peakData.pct, ...
    'VariableNames', {'peakName', 'peakChargeState', 'peakLoc', 'peakCounts', 'peakPct'});
if ~exist('peakConc','var'); peakConc = table(); end
peakConc = [peakConc; peakNew];
clearvars peakName peakChargeState

Repeat steps 1 and 3 for every peak the user wants. To compute a ratio between two elements, divide their peakCounts rows in peakConc.

Report

Show peakConc with each peak's name, charge state, location (Da), background-corrected counts, and percentage. Note that counts here are background-subtracted, so element ratios are more reliable than raw range counts.

Follow-ups:

• Export: writetable(peakConc, 'peakBackgroundCorrected.csv')
• For bulk background-removed composition over the whole spectrum, see posCalculateConcentrationBackgroundRemoved (note: in untested/).
• Related helpers: tofSpecBackgroundDetermination, backgroundCorrectedPeakCount.