Untargeted Analysis

Untargeted mode discovers metabolite features from your RAW files without needing a metabolite list up front. LEAF detects every peak it can, aligns them across samples, and gives you a feature table to triage.

[Screenshot: Untargeted view showing feature table and EIC chart]

When to use it

Use untargeted when...	Use targeted when...
You don't know what's in your sample	You have a defined panel of compounds to quantify
You're hunting for unknown metabolites	You're tracking specific pathways
You want a survey of one group vs another	You need quantitative comparisons of named compounds

You can always run both: untargeted to find candidates, targeted to lock them in for routine analysis.

Switch to untargeted mode

On the Extract page, click the Targeted / Untargeted toggle at the top. The compound list editor disappears (you don't need a CSV) and the parameters sidebar swaps in untargeted-specific options.

Parameters

Parameter	Default	What it does
Polarity	NEG	Match your method's polarity
Mass Tolerance	5 ppm	Tighter than targeted — affects feature alignment
Min intensity	1e5	Drop features below this peak height
Min samples	2	Require a feature to appear in at least N samples to keep it
RT range	full run	Optionally restrict to a part of the run

Run the extraction

Click Start Processing. Progress shows in the same floating action button as targeted runs. Untargeted runs typically take 2–5× longer than targeted because every peak has to be detected, not just the ones in your list.

Open the results

After completion, click Open in the jobs panel. The Untargeted view loads.

Layout

Panel	What it shows
Feature table	Every detected feature with m/z, RT, detection rate, intensity stats
EIC chart	Chromatogram for the selected feature across samples
Alignment panel	RT alignment quality across samples
Gap-group panel	Features that should align but don't — flagged for review
Stats panel	Per-feature group comparisons (fold change, p-value)
Results panel	Filter, sort, and tag features for export

Triage workflow

1. Filter the feature table by detection rate (e.g., keep only features in >50% of samples)
2. Sort by intensity or by group fold-change to find candidates
3. Inspect the EIC chart — does the peak look real, or is it noise?
4. Tag interesting features with flags or notes
5. Export the tagged set as a CSV — feed it back into LEAF as a targeted CSV for routine quantification

Identify features

LEAF doesn't identify features for you — it gives you m/z and RT. To get a name, search:

• A spectral library (e.g., HMDB, METLIN, MoNA)
• An MS2 spectrum from the same sample (LEAF supports MS2 matching against mzVault libraries — see the LEAF developer docs)
• A pure standard run on the same instrument

When you have a name, add the feature to a targeted CSV for the next batch of samples.

Export

Untargeted exports save as .usd files — same format family as .msd, but with feature data instead of compound data. The export dialog also offers per-feature CSVs.

→ Export details

Next step

→ UI tour — every panel and button explained