Prepare Data

Targeted LEAF analyses require two inputs: a folder of LC-MS files and a compound list. The compound list defines the ions or transitions to extract, while the data folder provides the chromatographic measurements.

Workflow

LC-MS files

LEAF reads targeted inputs from .raw, .mzml, and .mzml.gz files in the web UI and CLI. The Python extractor also supports Shimadzu .lcd files when the compound list supplies explicit MRM transitions. Put all files for one experiment in a single folder and use one acquisition format per extraction; mixed formats are rejected.

Naming: use descriptive filenames such as WT_rep1.raw, WT_rep2.raw, and KO_rep1.raw. LEAF can derive clean sample names from filenames and can group replicates by name prefix during review.

Blanks: include blank injections when they are part of the experiment. LEAF skips files with blank in the filename by default during targeted extraction.

Metabolite list (CSV / TSV)

A CSV or TSV lists the compounds or transitions to quantify. LEAF maps common exported formats into its internal schema based on column headers:

• LEAF native (RFA) — minimal columns described below
• Skyline transition / target lists — CSV/TSV files used for Skyline imports
• El-MAVEN compound databases or peak tables — El-MAVEN input files and El-MAVEN exported peak tables

Header matching is case-insensitive and ignores spaces and underscores. For example, precursor_name, Precursor Name, and PrecursorName map to the same field. Tab-delimited files are detected from the header line.

What LEAF needs

After column mapping, LEAF requires a compound name, a retention time, and either a molecular formula or an explicit transition. Formula-based targeted extraction is the common LC-MS metabolomics path; transition rows are used for MRM/SRM datasets such as Shimadzu .lcd inputs.

LEAF field	Required?	LEAF native	Skyline	El-MAVEN
Compound name	✓	Metabolite / Compound / Name	Precursor Name / Molecule Name	compoundName / compoundId
Retention time (min)	✓	RetentionTime / RT	Explicit Retention Time	medRt / expectedRt
Molecular formula	conditional	Formula	Precursor Formula / Molecular Formula / Molecule Formula	Formula
Transition	conditional	Transition	transition columns	—
Adduct	recommended	Adduct	Precursor Adduct / Precursor Charge	adductName / ionName / charge
Exact m/z	optional	Mass	Precursor m/z	medMz
Tracing group	optional	TracingGroup / Group	Molecule List Name	—

If Adduct is missing in a formula-based list, LEAF uses the default adduct for the selected polarity (M-H for NEG, M+H for POS). If Mass is missing, LEAF computes the expected m/z from formula and adduct.

El-MAVEN isotopologue rows

If an El-MAVEN peak table contains per-isotopologue rows (isotopeLabel or label values such as C12 PARENT, C13-label-1, or M+0), LEAF keeps the parent row per compound. Isotopologue extraction is configured separately in the Tracing Editor.

Example (LEAF native)

Metabolite,Formula,RetentionTime,Adduct
Glucose,C6H12O6,5.2,M-H
Lactate,C3H6O3,3.1,M-H
Citrate,C6H8O7,8.5,M-H
Pyruvate,C3H4O3,2.8,M-H
Glutamate,C5H9NO4,4.1,M-H
Succinate,C4H6O4,7.2,M-H
Malate,C4H6O5,6.5,M-H
Fumarate,C4H4O4,7.0,M-H

A starter list for primary metabolism is included with LEAF: examples/metabolite-list-primary-metabolism.csv.

For isotope tracing

Use the same compound CSV. Isotopologues (M+1, M+2, ...) are configured in the Tracing Editor on the Extract page rather than in the CSV. See Isotope tracing.

Validation before extraction

After uploading the CSV, click Validate in the compound list editor. LEAF flags:

• Missing required columns (Compound, Formula, RetentionTime)
• Empty compound names
• Duplicate compound + adduct combinations
• Invalid molecular formulas (typos, unsupported elements)
• Non-positive retention times

Resolve validation errors before starting extraction. Warnings should be reviewed because they may affect quantitative interpretation.

[Screenshot: compound list editor showing validation warnings]

Also from a script

The compound CSV is identical for the CLI and Python paths — no separate format. Pass it as the second positional argument to leaf targeted:

leaf targeted ./samples ./compounds.csv ./outputs

→ leaf targeted reference

Next step

→ Extract