Assembly does not always produce one clean, circular mitogenome. A single sample can yield several assembly paths, and a path can be broken into several scaffolds. MitoPilot keeps every retained path and scaffold rather than discarding them silently, and gives you tools to decide what each one represents before annotation and export.
What you see depends partly on the assembler. MitoPilot can assemble
with GetOrganelle
(the default) or MitoFinder, set by
the assembler option at project setup. GetOrganelle
explores an assembly graph and can return several alternative
paths through it, each possibly split into scaffolds.
MitoFinder returns a single path, but can still report
several scaffolds when it recovers more than one
mitochondrial contig. The strategies below apply the same way regardless
of which assembler produced the sequences.
The right strategy depends on what the extra sequences actually are, and the two cases have different biological meanings:
- Multiple paths are usually competing reconstructions of the same genome from one tangled assembly graph. You want a single answer, so you resolve them down to one.
- Multiple scaffolds in one path are pieces, and they can mean different things. They may be fragments of one genome that an incomplete assembly failed to connect; they may be genuinely separate molecules that belong together, as in a naturally multi-chromosome mitogenome; or they may be unrelated sequence, such as contamination or a NUMT. Fragments of the same mitogenome should be joined, while separate molecules should be kept apart.
Each retained (path, scaffold) combination becomes its
own row in the Annotate module and is annotated and validated
independently. The number of paths and scaffolds kept per sample is
capped by the Max assembly paths and Max
scaffolds assembly options (default 10 each); samples above
that threshold will not be allowed to proceed to annotation.
Resolving multiple paths
Multiple paths come only from GetOrganelle; MitoFinder always returns a single path. When a GetOrganelle sample assembles into more than one path, it is flagged in the Assemble module (“Unable to resolve single assembly from reads”). Open the assembly details to see each path. Because the paths are alternative hypotheses about one genome, you should reduce them to a single assembly before continuing. There are two ways to do this.
Ignore the extras. If one path is clearly the correct mitogenome, use the ignore buttons to drop the others. Only the remaining path advances to annotation.
Build a consensus (Path 0). When the paths disagree only in a few regions, use Trim to consensus to reconcile them into a single sequence. MitoPilot aligns the paths, and for each block where they conflict you choose which path to use (or mask with ambiguity codes). The result is saved as Path 0, a single edited assembly that supersedes the alternatives.
Either way, the sample arrives at annotation as one assembly.
Joining or separating scaffolds
A single path split across several scaffolds is handled in the assembly details as well. If the scaffolds are pieces of one fragmented mitogenome, join them; if they are separate molecules that genuinely coexist, keep them apart.
Join scaffolds (Path 0). For any single-path sample with more than one scaffold, the assembly details window offers a “Join scaffolds into one assembly” editor. The reference-guided layout (Re-map to reference) orders and orients the scaffolds against a chosen BLAST reference; you can override the order and orientation and mark the assembly circular, then Build joined assembly (Path 0). If the scaffolds carry different BLAST hits, MitoPilot warns you, because joining unrelated sequences produces poor overlaps and an unreliable assembly.
The Join scaffolds button in the assembly options window is a separate, optional convenience rather than a switch for the editor. When it is enabled, the pipeline builds Path 0 automatically during assembly for eligible samples, but only when their scaffolds map to the same reference, so you do not have to join each fragmented sample by hand. It is off by default. Leave it off when your samples may be naturally multipartite or contaminated (see below), so their scaffolds are kept separate rather than auto-joined.
Keep them separate. Not every fragmented sample should be joined. Some mitogenomes are naturally multipartite, with the genes distributed across several small circular or linear chromosomes, and those scaffolds are separate molecules of one real genome that must not be concatenated. Other times the scaffolds are unrelated: contamination, a NUMT, or distinct heteroplasmic variants. In all of these cases, leave the scaffolds separate. Each is annotated and validated on its own, so a single-copy gene present once per molecule is not mistaken for a duplicate, and each is exported as its own record.
Exporting
Export behaves differently for the two cases, matching the biology.
Multiple scaffolds export as separate records. A
sample contributing more than one scaffold produces one GenBank record
per scaffold. To keep SeqIDs unique within a submission, MitoPilot add
as suffix as ID_p<path>_s<scaffold> (a
single-unit sample keeps its plain ID). Because a sample
split across scaffolds could instead be one fragmented genome,
export asks you to confirm.
Multiple paths cannot be exported. Exporting every path would submit several near-identical records for one specimen, so MitoPilot blocks export of any sample that still has more than one assembly path and will emit a warning message. Resolve by ignoring the extra paths or building a consensus Path 0, then export.