Is it in it?
A Specified Search Run answers one question about your signal: does it contain what you are looking for — at the thresholds you set — and what is the evidence?
What you can search for
- Sequences: a FASTA of up to 50 targets in the base price (plasmids, genes or CDS, variants such as SNVs and indels, gene or strain markers, phage or viral genomes, insert or construct sequences)
- Constructs: did the insert integrate, is the plasmid in the strain, is the product the one you designed
- Loci on a reference: a reference FASTA plus a BED of regions
- Modification positions: is 5mC, 5hmC, 6mA or 4mC at these loci, and in what fraction of reads
- After an Alphabet Extension: your extended symbols
Optionally, a background reference to subtract known reads (host, medium, carrier).
Examples we like: is the phage in the fermenter run; is the target organism sequence in the environmental sample; did the construct land; is the mark at the locus.
The reference genome
Every Search Run is aligned against a reference you name, and we cannot start one without it. You can give it to us three ways:
- An NCBI assembly accession — GCF_ or GCA_, with the version after the dot. Nothing to upload, and it pins the exact build including its patch level.
- A FASTA (.fasta or .fa), uploaded from your job page over short-lived credentials that can write to nothing else.
- Optionally, a .bed of loci, genes or chromosomes to scope the search. Never required.
This is not paperwork. A BAM is reads placed against a reference, and the same run aligned to a different assembly gives different contig names and different coordinates. If we align to one build and you compare against another, the two disagree and nothing in either file says why. Naming the build is what makes our result and your own comparison the same measurement.
If there is no reference for your organism yet, a Search Run is not where to start — there is nothing to align to, and coordinates from an assembly built on the fly are not reproducible. Building one is development work: it begins with the feasibility stage of a Calibration, which ends in a written answer about whether your data supports an assembly at all.
You set the thresholds
We offer defaults; you can change them: minimum supporting reads (default 3), minimum read identity to target (default 90 %), minimum coverage breadth (default 50 %). For modification queries: minimum modified-read fraction and minimum depth at the locus. The thresholds go into your audit record.
The three calls
- Evidence foundreads meet your thresholds for this target, and the decoy set does not explain them.
- Not observed at this depthnothing met your thresholds. We say “at this depth” because a search has a depth, and absence is not something a search can claim. The number of reads examined is in the report.
- Inconclusivehits exist but sit below your thresholds, or are indistinguishable from decoy hits.
The decoy set
For every target we generate a decoy of the same length and composition (shuffled and reversed) and search for it too. The decoy hit rate tells you how many spurious matches your thresholds let through. It is printed next to every call.
What you receive
Search report (HTML and JSON) · evidence reads per target as FASTQ and BAM subset, aligned against the reference you named and recorded in your audit record — these are the only reads we basecall · .sen-lite reconstruction of the top evidence reads per target · signed audit record with the hash of your target set, your thresholds and your reference · viewer. Full FASTQ/BAM of the whole run: add-on.
Prices
Search Run or Decode Run? A Search Run answers whether your targets are in the signal, working on the raw current without basecalling the run. A Decode Run basecalls every read and returns the full sequence, modifications and reconstruction.
From €249 per run (up to 150 Gsamples, up to 50 targets) — from €11 per barcode when multiplexed. Full outputs +€199. Repeat search on a retained job €99. More targets +€49 per further 50. If you'd rather we build the target set from names or accessions: +€149.
Research use only — what we do and do not accept
Research observation at customer-specified thresholds; not a validated assay; no limit of detection is claimed.
This search is for research use only. No clinical, diagnostic, forensic, or regulatory decision depends on its result.
Patient specimens and sequences that serve identification of a natural person are not accepted.
Every first Search Run from a new customer is read by a person before we send an offer. If your description leaves a question, we ask one; if it is still unclear, we decline politely.
Try it for free
Send up to 5 Gsamples and up to three targets with your Signal Check.