Sequence annotation

A sequence file on its own tells you almost nothing. It is a long string with no landmarks, and the work of finding the landmarks — where the coding region starts, which frame it reads in, where the promoter sits, whether that stop codon is real or an artefact of the frame you happened to pick — is the work everyone does first and nobody enjoys doing twice.

The annotation engine does that pass for you. Built on Biopython, it parses what you give it, marks the features it can identify, scans every frame on both strands for open reading frames, and produces translations alongside the nucleotide view. The result is rendered as an interactive map through seqviz, linear or circular, zoomable from whole-plasmid down to the individual base.

It is the entry point to most of the platform. Sequences arrive here before they go anywhere else, because everything downstream — assembly, primer design, codon adaptation — works better against something that has already been described.

Upload a GenBank or FASTA file, paste a sequence directly, or pick one that is already in the system: a gene from any of the 65 organisms in the catalog, a construct from a project, or a record carried over from an earlier run. Catalog entries are addressable by their ID, so pulling the sequence behind SS-0421 is a lookup rather than a download-and-reupload round trip.

Existing annotation is respected. If your GenBank file already carries features, they come through intact and the engine adds to them rather than overwriting — a distinction that matters when the file has been curated by hand and you would rather not lose that work.

Engine

Biopython

Viewer

seqviz, linear and circular

Accepts

GenBank, FASTA, pasted sequence, catalog ID

Exports

GenBank, FASTA

Engine class

CPU

Credit cost

Free

Typical runtime

Seconds for a gene, under a minute for a plasmid

Annotation is a CPU engine, which on this platform means it costs nothing to run. There is no quota to watch and no reason to batch your work to save credits — annotate the same sequence six times with different filter settings if that is what it takes. Credits exist only for GPU engines, where $25 buys 250 credits and one credit covers roughly one Standard job.

Exports are GenBank and FASTA. GenBank carries the annotation with it, which is the point: the file you download opens in whatever desktop tool your collaborator prefers and still knows where its features are. FASTA is there for the cases where you just need the sequence and nothing else.

Researchers use it as a first look — a sequence arrives from a collaborator or a sequencing run and needs to be understood before anything is decided about it. Educators use the map as a teaching surface, because a zoomable rendering of a real construct explains reading frames faster than a diagram does. Developers pull the same annotation through the API and never open the viewer at all.

Inside the catalog, it is simply infrastructure. Every one of the 1,000 constructs and all 14.1 Mbp of designed DNA has been through this engine, which is why the records behind the heat ladder — from 100 SHU at the gentle end to 16,000,000 SHU at the top — read as annotated designs rather than as anonymous strings.