SeqBench

Whole-Plasmid Sequencing Check — Nanopore Consensus vs Map

Per-sample verdict on whole-plasmid sequencing results, against your map.

Your files are never stored or shared

Check whole-plasmid nanopore sequencing results against your plasmid map. Drop the map (SnapGene .dna or GenBank) and what your provider sent — the order's results .zip, or one consensus FASTA, GenBank or .ab1 — and every sample gets one verdict: use it, confirm by Sanger, look at it, or pick another colony. Assemblers start the circle at an arbitrary base and often return the reverse complement, so rotation and strand are corrected first; each difference is then reported in your map's numbering and classed as silent, missense or premature stop. Differences in a documented nanopore error context (a homopolymer over 9, a Dam or Dcm methylation site) ask for one Sanger read instead of a new colony, and the primer for it is planned for you.

0 bp. An annotated record is what turns a position into “silent in AmpR” or “a stop at codon 41”.

No map to hand? Name the plasmid you built into it, or give its GenBank accession.

A name is read from the site's vector library; an accession is fetched from NCBI, UniProt or Ensembl, features included. Pressing the button is what sends it — nothing leaves the page while you type.

0 bp

Design, edit, and verify complete constructsIt opens on the verification panel, where reads are compared with it.

How to use the Whole-Plasmid Sequencing Check tool

  1. 1Drop or paste your map (.dna or GenBank), then the provider's results .zip or one consensus file. An order gives one verdict per sample.
  2. 2Read the decision, then the reasons under it: each names the observation it rests on.
  3. 3Add the reads (.fastq) if you have them — read lengths, the reads behind each difference, and a pileup to click into.
  4. 4Take the clone away as .dna, .gb or FASTA: your map with this consensus's differences written in. Artifact-context differences are left out and said so.

FAQ

My consensus and my map start at different bases. Does that matter?

No. Assemblers start a circular consensus at an arbitrary base, so compared end to end the two look unrelated. The rotation is found automatically, and the result reports the offset and whether the consensus was reverse-complemented.

Is a 1 bp indel a real mutation or a sequencing error?

The sequence alone cannot tell, which is why three different homopolymer thresholds are reported and none of them is a verdict. A run of 4 or more is counted as slippage context (an indel there is what a sequencing reaction and a real slippage mutation both look like); a run of 7 or more is flagged by the provider itself; and 10 or more is the provider's own documented nanopore error mode. The Reads column, when the results .zip carried a per-base table, says how many reads actually carried the change.

What does the Reads column mean?

How many of the provider's reads show your map's version of each difference, out of how many cover that position, read off the per-base table in the results .zip. A substitution is exact. An indel is a lower bound: the table says a read had extra bases at a position, not which, so the count spans the run. A long indel gets coverage only. Reads agreeing is not proof — a methylation-site error appears in most reads alike — so the sentence says so where it applies.

Which differences should I confirm by Sanger before trusting them?

Those flagged as a known nanopore error context: a homopolymer of 10 or more, or an E. coli Dam (GATC) or Dcm (CCWGG) methylation site. Confirm these by Sanger before discarding a clone.

What do the reads add, and what do they need?

With the FASTQ from the same run: a read-length distribution against your plasmid's length (the answer to "the size looks wrong" or "two peaks"), the number of reads whose own alignment carries each difference split by strand, split-read rearrangements, and a pileup window at every difference. They are aligned server-side with minimap2. They are REPORTED, never used to change the verdict: the thresholds that would decide "two molecules" were measured on the provider's per-base tables, not on this pileup, so a sub-population call still comes from the consensus table.

What does this NOT check?

Coverage and mixed populations need the per-base read table, which is in the provider's results .zip but not in a consensus FASTA or GenBank file. From the .zip, read depth, mismatches and deletions are plotted and the sample is checked for a second molecule. For raw reads, use Sequencing Read-Back Verification.

Helpful?