Data Format Reference

Feature Quantification Table

The feature table must be a comma-separated (.csv) file with at minimum these columns:

Column name	Example	Notes
row ID	1, 2, 3	Integer, must match SCANS= in MGF
row m/z	256.0843	Precursor m/z
row retention time	1.245	In minutes
Sample1.mzML Peak area	45231	One column per sample, exactly this format

First three rows example:

row ID,row m/z,row retention time,Sample1.mzML Peak area,Sample2.mzML Peak area
1,256.0843,1.245,45231,12450
2,431.2012,3.812,0,98320
3,189.0558,2.104,22100,22050

Common mistakes:

• Column headers must be exact (case-sensitive). Row ID ≠ row ID.
• Sample columns must end in ` Peak area` (with a space before “Peak”).
• Empty intensity cells should be 0, not blank.

---

Metadata Table

The metadata file must be a tab-separated (.tsv) file.

Column name	Example	Notes
filename	Sample1.mzML	Must exactly match feature table sample names
ATTRIBUTE_Group	treatment	Any column starting with ATTRIBUTE_ becomes a group

First three rows example:

filename	ATTRIBUTE_Group	ATTRIBUTE_Timepoint
Sample1.mzML	treatment	early
Sample2.mzML	control	early
Sample3.mzML	treatment	late

Common mistakes:

• The filename column values must exactly match (including extension and case) what appears in the feature table sample column headers.
• Do not use blank cells. Use N_A (with an underscore) for missing values.

---

MS/MS Spectra (MGF)

Your MGF file should be exported directly from your feature-finding software (MZmine “Export feature list as MGF”, XCMS, etc.). Each spectrum block must include:

BEGIN IONS
SCANS=1
PEPMASS=256.0843
CHARGE=1+
100.2 5432.1
150.3 8921.0
END IONS

The SCANS= value must match the row ID in your feature table.

If your software uses FEATURE_ID= instead of SCANS=, this is handled automatically for MZmine3 output.