_FUNCnnn.STS

Binary scan statistics file. One file per function (e.g. _FUNC001.STS, _FUNC002.STS). Records per-scan instrument housekeeping values: electrode voltages, collision energies, push counts, lock-mass corrections, and TIC traces.

Status: Fully Decoded

File Layout

[32-byte file preamble]
[n_desc × 48-byte descriptor records]      (starting at offset 0x20)
[n_scans × scan_record_size bytes of data]

Total header size = 32 + n_desc × 48. This equals the data_offset stored at preamble byte 0.

File Preamble (32 bytes)

Offset	Type	Description
0x00	u16	data_offset - byte offset where per-scan data begins
0x02	u16	Format version (always 1 observed)
0x04	u16	scan_record_size - bytes per scan in the data section
0x06	u16	n_desc - number of descriptor records
0x08-0x1F	-	Zeroes (padding to 32 bytes)

Equations:

• data_offset = 32 + n_desc * 48
• n_scans = (file_size - data_offset) / scan_record_size

Validated:

• PXD058812 P15: data_offset=0x4D0=1232, scan_sz=63, n_desc=25, n_scans=197
• PXD075602 DHPR: data_offset=0x0AA0=2720, scan_sz=167, n_desc=56, n_scans=1150

Descriptor Records (48 bytes each)

Each descriptor defines one channel stored in the per-scan data section.

Offset within record	Type	Description
0x00	u16	Channel sequence number (not always contiguous; some skip)
0x02	u16	Encoding type (see table below)
0x04	u16	Byte offset of this channel within each scan record
0x06-0x2F	bytes	Null-padded ASCII channel name (42 bytes)

Encoding Types

Type code	C type	Size	Notes
0	u8	1 byte	Boolean flag / small integer
1	i16	2 bytes	Signed 16-bit integer
2	u32	4 bytes	Unsigned 32-bit integer
3	f32	4 bytes	IEEE 754 single-precision float

The byte offset in the descriptor (0x04) is used directly as an index into the per-scan record. Channels are packed without explicit gaps; the offset of channel k+1 equals offset k + size(type k).

Per-Scan Data

The data section starts at data_offset and contains n_scans consecutive records, each scan_record_size bytes. To read channel X for scan S:

channel_value = scan_data[S * scan_record_size + descriptor[X].offset]

decoded using the descriptor's encoding type.

Known Channels

Observed across all four corpus datasets. Channels available depend on instrument generation; newer firmware adds channels.

Seq	Name	Type	Code	Notes
13	Segment Number	i16	varies	0 for normal scans
52	Cone	i16	varies	Cone voltage (V); matches _extern.inf Sampling Cone
62	Collision Energy	f32	varies	Per-scan collision energy (eV)
76/78	Pusher Frequency / Collision Energy2	varies	varies	Function-dependent
77	Set Mass	f32	3	Precursor m/z of an MS/MS scan; 0 on MS1, MSe and lock-mass scans
100	Reference Scan	u8	varies	1 if this scan is a lock-mass reference
101	Use lock mass correction	u8	varies	Boolean
102	Lock mass correction	f32	varies	Mass correction applied (Da)
103	Use temp correction	u8	varies	Boolean
104	Temperature correction	f32	varies	Temperature-derived TOF correction (µs)
107	TIC Trace A	f32	varies	Total Ion Current (a.u.) for function A
108	TIC Trace B	f32	varies	Total Ion Current (a.u.) for function B
116	Scan Push Count	u32	varies	Number of ADC pushes summed in this scan
117	Scanwave Normalisation Factor	f32	varies	IMS normalisation factor
121	ETD Fragmentation Mode	i16	varies	0 = CID; non-zero = ETD

Not all channels are present in every file. Parse by looking up the descriptor list rather than assuming fixed offsets.

Set Mass (seq 77)

In DDA bundles the _extern.inf Set Mass of each MS/MS function is 0 or a fixed method default, while this channel holds each scan's precursor m/z. Every MS/MS scan of every DDA function in the corpus has a non-zero value; every MS1, MSe and lock-mass scan has 0.

Cross-check against _auto.txt, a plain-text m/z,time (s) log of the precursors chosen during acquisition, present in 6 corpus bundles. Each run of consecutive MS/MS scans with one Set Mass value starting within 30 s after a log entry was matched to it:

Bundle	_auto.txt entries	MS/MS runs	Matched
PXD041695/JLG-1D-ISY-C18_Q0764.raw	410	410	410 within 0.5 mDa
PXD041695/JLG-1D-ISY-C8_Q0753.raw	412	412	412 within 0.5 mDa
PXD010569/20161025_AV_Dario_Banda_10.raw	274	274	274 within 0.5 mDa
PXD010569/20161025_AV_Dario_Banda_3.raw	273	273	273 within 0.5 mDa
PXD001123/SiRNA_21-2.raw	882	865	865 within 0.5 mDa
PXD029515/blast_young_0h_H1__MSMS.raw	1217	1217	1217 within 100 ppm (median +18 ppm)

In PXD001123, 17 logged precursors have no MS/MS scan. In PXD029515 the log and the channel differ by a constant offset of about 18 ppm; the channel value is the one reported. PXD003126 and PXD073126 (mobility DDA) have no _auto.txt, but their MS/MS scans carry the channel the same way. For the targeted MS/MS sample PXD035818 the channel reads 884.92 on every scan, where _extern.inf gives Set Mass 884.9.

Observed Values

Dataset	n_desc	scan_sz	Cone	Collision Energy	Push Count
PXD058812 P15 (non-IMS QTOF)	25	63	30 V	10.0 eV	-
PXD075602 DHPR (Xevo G2-XS)	56	167	30 V	4.0 eV	8299

Push Count = 8299 in DHPR: 8299 × 60.25 µs ≈ 0.50 s per scan, consistent with the 0.50 s scan time recorded in _FUNCTNS.INF.

Relationship to IDX Fields

The IDX +0x08 field (30-byte Variant B, value labelled "Large value / TIC?") likely corresponds to one of the TIC Trace channels here. This cross-check has not yet been confirmed empirically.

Rust Implementation

crate::raw::func_sts::FuncSts parses this file and exposes per-scan channel lookups by name; FuncSts::collision_energy(scan_idx) reads the "Collision Energy" channel specifically, and FuncSts::etd_fragmentation_mode(scan_idx) reads the "ETD Fragmentation Mode" channel (seq 121) specifically, and FuncSts::set_mass(scan_idx) reads the "Set Mass" channel (seq 77), with a stored 0 returned as None. Reader::precursor_mz uses it as the precursor m/z of each scan, falling back to the _extern.inf Set Mass (also None when 0), and the mzML records write that value as both the isolation window target m/z and the selected ion m/z. Reader::open parses every function's _FUNCnnn.STS when present (silently None if absent or malformed - this file is supplementary, not required to decode peak data), and mzml::precursor_info_for uses it to populate PrecursorInfo::collision_energy for every MS2 function, including MSe/HDMSe (which has no discrete precursor target_mz but still reports a real per-scan collision energy) and targeted MS/MS (Sigilweaver/OpenWRaw#8, #13), and to populate PrecursorInfo::activation from the ETD Fragmentation Mode channel: 0 maps to Activation::CID, non-zero to Activation::ETD (Sigilweaver/OpenWRaw#22). Every ETD Fragmentation Mode value seen in the repository corpus is 0; the non-zero branch follows directly from this table's documented encoding but has not been exercised against a real non-zero fixture, only synthetically (mzml::tests::activation_for_maps_zero_to_cid_and_nonzero_to_etd).

Channels 101 and 102 remain available through the generic channel / value_at API but are deliberately not applied downstream. Every sample in the repository corpus has Use lock mass correction = 0 and Lock mass correction = 0.0. Consequently the corpus cannot establish how a non-zero lock-mass value modifies the calibrated TOF-to-m/z conversion. Applying an interpretation would require a fixture with a non-zero value and an independently checkable in-repository invariant (Sigilweaver/OpenWRaw#23).

Reference Sources

• Empirical analysis: all _FUNC001.STS files in corpus
• Corpus samples:
  • PXD058812/molecular_mass_P15_01.raw (25 channels, 63 bytes/scan)
  • PXD075602/DHPR_11257-1.raw (56 channels, 167 bytes/scan)
  • PXD035818/17122018_TNFA_PEPTIDE_GSHH_MSMS_884.raw (targeted MS/MS, confirms the same descriptor/channel layout on a non-MSe function)