@HD VN:1.6 SO:unknown
@PG ID:basecaller PN:dorado VN:0.2.4+3fc2b0f CL:dorado basecaller dna_r10.4.1_e8.2_400bps_hac@v4.1.0 pod5/ DS:gpu:Quadro GV100
| RG | ID | <runid>_<basecalling_model>_<barcode_arrangement> |
| PU | <flow_cell_id> |
|
| PM | <device_id> |
|
| DT | <exp_start_time> |
|
| PL | ONT |
|
| DS | basecall_model=<basecall_model_name> modbase_models=<modbase_model_names> runid=<run_id> |
|
| LB | <sample_id> |
|
| SM | <sample_id> |
| RG:Z: | <runid>_<basecalling_model>_<barcode_arrangement> |
| qs:f: | mean basecall qscore |
| ts:i: | the number of samples trimmed from the start of the signal |
| ns:i: | the basecalled sequence corresponds to the interval signal[ts : ns] the move table maps to the same interval. note that ns reflects trimming (if any) from the rear of the signal. |
| mx:i: | read mux |
| ch:i: | read channel |
| rn:i: | read number |
| st:Z: | read start time (in UTC) |
| du:f: | duration of the read (in seconds) |
| fn:Z: | file name |
| sm:f: | scaling midpoint/mean/median (pA to ~0-mean/1-sd) |
| sd:f: | scaling dispersion (pA to ~0-mean/1-sd) |
| sv:Z: | scaling version |
| mv:B:c | sequence to signal move table (optional) |
| dx:i: | bool to signify duplex read (only in duplex mode) |
| pi:Z: | parent read id for a split read |
| sp:i: | start coordinate of split read in parent read signal |
| pt:i: | estimated poly(A/T) tail length in cDNA and dRNA reads |
| bh:i: | number of detected bedfile hits (only if alignment was performed with a specified bed-file) |
| MN:i: | Length of sequence at the time MM and ML were produced |
When modified base output is requested (via the --modified-bases CLI argument), the modified base calls will be output directly in the output files via SAM tags.
The MM and ML tags are specified in the SAM format specification documentation.
Briefly, these tags represent the relative positions and probability that particular canonical bases have the specified modified bases.
These tags in the SAM/BAM/CRAM formats can be parsed by the modkit software for downstream analysis.
For aligned outputs, visualization of these tags is available in popular genome browsers, including IGV and JBrowse.
When dorado is run with alignment enabled, additional tags from minimap2 are added to each SAM record. Details of those tags
are available on the minimap2 manpage.
When a single input read contains multiple concatenated reads, dorado basecaller will split the original input read into separate subreads. This operation is performed by default for both DNA and RNA. Each subread has a new read id that is assigned by dorado. The following tags can be used to associate a subread to its parent:
pi:Zcontains the parent read id the subread was generated from.sp:imaps the start of the subread's signal data to the corresponding location in the parent read's signal data.ns:iis the number of samples corresponding to the subread after splitting.ts:iis the number samples trimmed from the start of subread's signal after splitting.