#The run

#Traces and dyes

Each dye has a raw trace, the detector signal per scan, and an analysed one, after the instrument software's baseline and spectral correction. rawTrace and analysedTrace take the dye number and know where it is stored: dyes 1 to 4 in DATA 1 to 4 and 9 to 12, a fifth and later dye in DATA 105 and up and 205 and up. DATA 5 to 8 are instrument readings (voltage, current, power, temperature), not dyes.

for (def d in abif.dyes($f)) {
    def trace as list of int init abif.rawTrace($f, $d.number);
    io.printf("%s at %d nm: %d scans\n", $d.name, $d.wavelength, len($trace));
}

dyes counts the dyes from Dye# (or the DyeN tags when it is missing) and reads each name and wavelength; one the file lacks, or does not store as text or one integer (a HID file encrypts most names), is "" or 0.

#Base calls

A sequencing run (.ab1) carries the basecaller's result: the bases (PBAS), a quality value per base (PCON) and the scan where each base's peak is (PLOC). calls reads them as called (number 2), editedCalls as a user edited them (number 1); in an unedited file the two are equal.

def c as abif.Calls init abif.calls($f);
for (def i in 0..len($c.qualities)) {
    if ($c.qualities[$i] < 20) {
        io.printf("weak call %s at %d\n", $c.bases[$i..$i + 1], $i + 1);
    }
}

A fragment run has no base calls, and calls says so. Qualities or peaks a file lacks are an empty list; when they are there, they must have one entry per base.

#Metadata

info collects the sample and run metadata into one Info: sample name, well, lane, instrument, model, run name and module, dye set, polymer, size standard, and the start and end of the run and of data collection as YYYY-MM-DD HH:MM:SS.

def i as abif.Info init abif.info($f);
io.printf("%s in %s on %s, %s\n", $i.sample, $i.well, $i.model, $i.runStart);

Fixed-width fields are padded (MODL is "310 ", GTyp "POP7" and 28 spaces), so info removes trailing spaces and NULs; text returns the stored value. A field the file lacks, or keeps encrypted, is "".

#3500 HID files

A .hid file is ABIF version 3.00. Its sample and run metadata read as usual, but the traces, the peak tables and most dye names are stored as element types 28 to 34, which the published format does not define. Their bytes look random, their sizes are whole 8-byte blocks, and equal plaintexts give equal blocks: they are encrypted. tag returns them as Raw, the trace readers refuse them, and dyes and info leave their names "". Reading the signal of a HID run needs Applied Biosystems' software.