Fragment View¶
The fragment view holds what a DIA tool measured of each fragment of a precursor: its quantity, peak shape and, where a tool has it, the extracted ion chromatogram. It is optional, can be large, and is written only on request (--fragments). Specification 1.3 added it.
File: {PREFIX}.fragment.parquet. One row per fragment of a feature, channel and run. Identity composite: feature_id, label, ion_type, fragment_charge, neutral_loss, fragment_index.
| Field | Description | Type | Required |
|---|---|---|---|
fragment_id |
Opaque identity, derived from the composite | int64 | Yes |
feature_id |
The feature (feature.feature_id) the fragment belongs to; qpxc validate checks it |
int64 | Yes |
run_file_name, peptidoform, charge |
The precursor and run, repeated for reading without a join | string, string, int16 | Yes |
label |
Channel | string, null | No |
fragment_index |
1-based position in the tool's list, where it names fragments by position | int32, null | No |
ion_type |
Ion series and number, such as y8; null where the tool does not name the fragment |
string, null | No |
fragment_charge |
Charge of the fragment | int32, null | No |
neutral_loss |
Neutral loss such as H2O; null for none |
string, null | No |
mz / observed_mz |
Theoretical and measured m/z | float64, null | No |
intensity |
Raw quantity; null for not measured | float32, null | No |
corrected_intensity |
Quantity after the tool's correction | float32, null | No |
peak_height |
Peak height | float32, null | No |
correlation |
Correlation of the fragment's elution with the precursor's other fragments | float32, null | No |
excluded |
Whether the tool left the fragment out of the precursor's quantity | bool, null | No |
rt_array, intensity_array |
Extracted ion chromatogram, seconds and intensities. No converter fills them yet: they are for a producer that has them, as OpenMS' XIC files | list\<float32>, null | No |
additional_scores |
Other per-fragment values | list\<score>, null | No |
qpxc transform relayout can store additional_scores as a struct with one field per name (Struct lists).
Producers¶
DIA-NN 1.x --fragments |
DIA-NN 2.x --fragments |
Spectronaut --fragments |
|
|---|---|---|---|
fragment_index |
Position in Fragment.Quant.Raw |
Fr.N.Index + 1 |
--- |
ion_type, fragment_charge, neutral_loss, mz |
Fragment.Info, where the report has it; else null, and the position is all that names the fragment |
Fr.N.Id |
F.FrgIon, F.Charge, F.FrgLossType, F.TheoreticalMz |
intensity |
Fragment.Quant.Raw |
--- | F.PeakArea |
corrected_intensity |
Fragment.Quant.Corrected |
--- | F.NormalizedPeakArea |
correlation |
Fragment.Correlations |
--- | --- |
peak_height |
--- | Fr.N.Quantity |
F.PeakHeight |
observed_mz, excluded |
--- | --- | F.MeasuredMz, F.ExcludedFromQuantification |
additional_scores |
--- | Fr.N.Score as diann_fragment_score |
F.InterferenceScore, F.Log10SignalToNoise, F.MeasuredRelativeIntensity, F.PredictedRelativeIntensity, F.CalibratedMassAccuracy (PPM) |
DIA-NN 1.x lists every library fragment, with 0 for one it did not measure. DIA-NN 2.x writes them only when run with --export-quant, as one Fr.N.Index, Fr.N.Id, Fr.N.Quantity and Fr.N.Score group of columns per fragment: Fr.N.Index is the 0-based position in the library's fragment list, Fr.N.Quantity the fragment's peak height, not normalised, and Fr.N.Score the quality of its extracted elution profile; an empty Fr.N.Id is an unused slot. A fragment with no quantity has no row, so fragment_index can skip values.
Targeted runs¶
A transition of a PRM or SRM run is a fragment row. See Extensions.