Test project: 1. Assemble 2. Annotate 3. Export
The Export module turns finished annotations into the files GenBank needs for submission: a FASTA of the sequences and a five-column feature table of the annotations. MitoPilot allows you to export entire mitogenomes, individual genes, or both. Export also generates GFF3 annotation files and alignment summaries.
Make a group
Export works on groups, which correspond to submission batches. Select the samples you want and click ASSIGN GROUP.
The window summarizes what you selected: how many records, which topologies, and the gene order of each. Type a name (letters, numbers, dashes, and underscores only) and click Create.
For the test project, select everything. MitoPilot then stops you:
A GenBank submission may contain complete mitogenomes or partial
mitogenomes, but not both. This selection has some of each.
Split into two groups produces
fish_batch-complete and fish_batch-partial.
Note that completeness is derived from each assembly’s topology, with
circular meaning complete and linear meaning partial, unless you
overrode it using the per-sample Partial flag or the
linear_complete project setting.
A unit can only belong to one export group at a time. CLEAR GROUP removes the selected samples from their group. Re-assigning a grouped sample asks for confirmation first, because files already written for the old group are not cleaned up.
Export
Click EXPORT DATA.
Header Template is the FASTA definition line. Any
extra columns you put in the mapping file at project setup are available
here and can be referenced using {column_name}. This is how
you can include things like voucher numbers and BioSample accessions in
the GenBank submission files.
There are two special parameters. First, {completeness}
expands to “complete genome” or “partial genome” and belongs at the end
of the header for GenBank to parse it correctly. Second,
{seqid} takes the value in the ID column of the mapping
file and appends the scaffold/path number if that sample is exporting
multiple sequences.
Templates are validated as you type and can be saved by name for reuse.
Other options in the Export Data window:
- Generate Group-level PCG alignment summary aligns every protein-coding gene across the group and writes an HTML report. Worth the wait as a final check.
- Export individual protein-coding and rRNA genes writes per-gene FASTA files and feature tables, with their own header template. Useful for phylogenetics or single-gene submissions.
- Review PCG annotations for outliers runs a final pre-submission check described below.
Then click Export.
Note. If any sample in the group contributes more
than one record, MitoPilot asks you to confirm before exporting. Each
scaffold becomes its own GenBank record with a
ID_p<path>_s<scaffold> SeqID. This allows you
to submit samples that have truly fragmented mitogenomes. Samples that
still have more than one assembly path cannot be exported at all, since
different assembly paths usually represent multiple versions of the same
molecule.
The outlier review
With the review enabled, MitoPilot aligns each protein-coding gene across the group and stops on any gene where a sample looks out of line.
A sample’s gene is flagged when it appears truncated or extended
compared to the rest of the submission group, as determined by the
offset parameter settings.
Sequences are also flagged when their identity compared the rest of the group falls below the specified threshold. In the screenshot above, SRR21844202’s ATP6 is at 59.2% identity against the others.
You can page through the flagged genes with Prev and Next, jump straight into the annotation editor for any sample with edit, mark a gene resolved once you have looked at it, or cancel the export entirely. You can also edit genes for samples that were not automatically flagged.
Files are written only when you click Done, so edits made during review will appear in the exported files.
Tip. Low identity is not automatically an error. ATP8 and NAD6 are genuinely fast-evolving and routinely flag in a diverse export group.
Exported files
During export, files are saved to
<project>/out/export/<group name>/:
| File | What it is |
|---|---|
<group>.fasta |
The mitogenome sequences with your header template applied |
<group>.tbl |
The five-column feature table of annotations |
<group>_sample_info.csv |
The sample metadata behind the submission |
AA_alignments_<group>.html |
The group-level protein alignment report |
GFFs/ |
One GFF3 per sample, can be reviewed using Geneious or other tools |
genes/ |
Per-gene FASTA and feature tables, if you asked for them |
The FASTA and the feature table are the two files you need for a GenBank organelle submission.
The HTML report is the file to read first.
It opens with every flagged outlier in one table, then shows the alignment for each protein-coding gene across the group. Poor annotations tend to be obvious here.
Sequences containing unknown bases
Where MitoPilot joins a fragmented assembly, it inserts
Ns between the pieces to represent the gap. NCBI accepts
gapped mitogenome submissions, on one condition: the number of
Ns must be the estimated gap length. MitoPilot
therefore only joins across a junction whose size it can estimate from
the reference. A sample with a junction it cannot size is left
fragmented, with a note saying so; its contigs can be submitted as
several sequences under one BioSample, which NCBI also accepts.
On export, every run of Ns that MitoPilot inserted is
written as a gap feature carrying its estimated length,
however short the run:
6009 6020 gap
estimated_length 12That is all NCBI needs. The fuller assembly_gap feature
belongs to whole-genome assembly submissions and is not used here.
A feature spanning such a gap stays a single continuous
feature, since the gap length is an estimate rather than
unknown. A protein-coding gene also carries a note saying how many bases
are unknown, for example
contains 12 bases of unknown sequence, and codons the gap
leaves unreadable translate to X.
Runs of Ns that were already in a sequence you supplied
are exported as they are but are not declared as gaps:
they may be ambiguous base calls rather than a gap, and MitoPilot has
nothing to say about them. They still contribute the note on any coding
feature that contains them.
Watch the export messages for two things. A declared gap longer than 1000 bp is over the INSDC limit, and GenBank will query it. A feature whose start or end falls inside a gap is rejected outright: move the boundary in the annotation editor so it stops at the edge of the gap.
HYDRA Getting files off the cluster. Export runs inside the app rather than as a cluster job. You will need to download the data to submit to GenBank.
Congratulations, you’ve completed the test project walkthrough! When you are ready to try MitoPilot with your own data, see starting your own project.