Check a company-provided FASTA sequence against your expected PCR product — works fully offline in your browser.
Sequence verification means checking that the sequence a sequencing provider returned actually matches the PCR product you expected, and this tool does that comparison and issues a verdict you can act on: PASS, WARNING or SUSPICIOUS. It aligns the returned FASTA against your expected amplicon on both strands, calculates percent identity and coverage, counts mismatches, insertions and deletions, checks that your primers sit where they should, and reports the probability that a match this good happened by chance.
The thresholds are stated up front rather than hidden behind a score. Identity of 99.5% or above is a PASS. Between 95% and 99.5% is a WARNING, because the returned text is close but not identical and the differences deserve a look. Between 85% and 95% is also a WARNING, because the read may not be the right amplicon. Below 85% is SUSPICIOUS, meaning the sequence is almost certainly not derived from your product. Separately, coverage under 90% is flagged as an incomplete read, a read under 100 bases is called out as too short to verify anything, and a high fraction of ambiguous bases is reported as a quality problem rather than a mismatch.
This tool answers a specific and uncomfortable question: the sequencing company says the result is fine, and you are not sure whether to believe them. You paste in the FASTA they returned and the sequence you expected, and the report does the comparison, writes out the reasoning behind the verdict, and gives you something concrete to send back when you need to query the result.
It is built for the routine reality of small-scale sequencing work. Every construct, primer, gene or PCR product you order comes back as a text file, and a clean-looking result can still be wrong: the wrong clone, a PCR artefact, a carryover contaminant, or simply the wrong tube. Reading the text carefully by eye is slow and easy to get wrong, so this automates the comparison.
The report is generated entirely in your browser. Sequences that describe unpublished constructs, patent work or anything under an embargo never leave your device.
Align the returned FASTA against your expected product so mismatches, insertions and deletions are visible as evidence rather than as a hunch.
Compute sequence identity with a transparent threshold, so the verdict follows from a stated rule you can check and argue with rather than from an opaque score.
Get an explicit outcome with the evidence behind it, suitable for recording in a lab notebook or raising with the provider.
Download the report so the comparison, the numbers and the verdict are documented at the time the result arrived rather than reconstructed later.
Surface signs of a mixed or contaminated sample, which is a common cause of a result that looks almost right but is not usable.
Align the returned read against your expected amplicon on the forward strand and on its reverse complement, then score whichever matched better. Providers often return a read in the opposite orientation to the design, so this prevents a false mismatch report on a good result, and the report names the strand it used.
Optionally add your forward and reverse primers. The tool confirms the forward primer sits at the start of your expected amplicon, that the reverse primer reverse-complements to its end, and that the forward primer is actually present in the returned read, which is independent evidence that the read came from your reaction.
Every alignment is reported with a binomial tail probability: with 0.25 as the chance of any given base, the report calculates how improbable it is that a random sequence would match this well over the same number of aligned columns. A very small value is what turns a high identity percentage into real evidence rather than coincidence.
Coverage is the aligned length divided by your expected amplicon, and anything under 90 percent is flagged as an incomplete read. Reads under 100 bases are called out as too short to verify a PCR product, and the fraction of ambiguous IUPAC codes is reported so a low-quality read is not mistaken for a wrong clone.
Run a Needleman-Wunsch style global alignment for a full-length comparison, or a local alignment that finds the best-matching segment when a read only partly covers the amplicon. The scoring scheme and the counts of mismatches, insertions and deletions are printed so the alignment can be checked by eye.
The report states each conclusion in words: how the identity compares with the thresholds, whether coverage is complete, what the structural differences mean, whether the match could be coincidence, and what the primer evidence shows. A number you cannot interrogate is not evidence you can defend in a lab notebook or to a provider.
No. The alignment and scoring run in your browser, so unpublished constructs, patent work and anything under embargo stay on your machine.
PASS means the returned sequence matches your expectation within the stated tolerance. WARNING means there is a real discrepancy worth reviewing, often a single substitution that may be genuine or a sequencing artefact. SUSPICIOUS means the result diverges enough that you should not use it and should contact the provider.
No, and this is the most important caveat. A FASTA is only text. A perfect match proves the text agrees with your reference, not that the sequence came from your tube. The report says so explicitly, and it explains what evidence to request if provenance matters.
The raw AB1 chromatogram for Sanger sequencing, or the FASTQ for NGS, so you can inspect base calls and quality scores yourself. A provider unwilling to supply raw files is a warning sign, though not proof of anything.
Send the same PCR product to a second independent company and compare the results. Keep an aliquot so this is possible. Independent replication is the only check that actually addresses the provenance question.
A single-base substitution in an otherwise clean result is often just a sequencing artefact, so treat it as a warning rather than a failure. Repeated changes, clustered mismatches, unexpected insertions or deletions, or a length that differs from the expected amplicon are all stronger grounds to query the result.
This tool uses three thresholds: 99.5% identity or higher is a PASS, 95% to 99.5% is a WARNING because the text is close but not identical, 85% to 95% is a WARNING because the read may not be the right amplicon, and anything below 85% is SUSPICIOUS. Those are deliberately conservative defaults. Most vendors quote 98% or above as acceptable for a clean Sanger read, so treat the WARNING band as a prompt to read the differences table rather than as a failure.
Paste the FASTA the company returned and the PCR amplicon you expected, add your forward and reverse primers if you have them, and run the comparison. The tool aligns the returned read against your expected product on both strands, reports percent identity, coverage, mismatch, insertion and deletion counts, checks that your primers sit where they should, and returns a PASS, WARNING or SUSPICIOUS verdict with the reasoning written out.
Sanger sequencing is not perfectly accurate at the end of a read, and calling quality falls off after roughly 700 to 900 bases, so isolated mismatches are common even on a good sample. One or two substitutions in an otherwise clean alignment is usually a calling artefact. The report separates these from the cases that matter by also counting coverage and by flagging reads that fall below 90 percent of your expected amplicon.
N is an ambiguous base the sequencer could not call confidently, and the tool counts it along with the other IUPAC ambiguity codes R, Y, M, S, W, K, D, H, B and V. A high fraction of ambiguous bases means the read is low quality or incomplete rather than that your clone is wrong, so the report raises it as a quality flag separately from the identity verdict.
Yes. The tool aligns the returned sequence against your expected amplicon on both the forward strand and the reverse complement, then reports which strand matched and scores the better of the two. Sequencing providers frequently return a read in the opposite orientation to the way the construct was designed, so this avoids a false mismatch report on a perfectly good result.
Plain FASTA files with .fasta, .fa or .txt extensions, loaded from disk or pasted straight into the form. Header lines beginning with > are stripped and the sequence rows are concatenated, and bases are upper-cased with anything that is not a nucleotide or an IUPAC ambiguity code removed. You can also paste the sequences directly if you only have them in a browser or an email.
It checks a returned sequence against an expected amplicon, so it applies to any result that arrives as sequence text, including a consensus or assembled sequence from NGS. What it cannot do is assess raw NGS reads, because it does not read FASTQ with per-base quality scores. Use the sequence file converter and analyzer for FASTQ work, and this tool for verifying a final assembled sequence against what you expected.
Coverage is reported as the aligned length divided by your expected amplicon length, and anything under 90 percent is flagged. A short read weakens a PASS because part of your product was simply never examined, so the report tells you to request the full read. It also means the amplicon may be longer than the read window your provider used.
Both primer fields are optional, but adding them lets the tool check that the forward primer sits at the start of your expected amplicon, that the reverse primer reverse-complements to its end, and that the forward primer is actually present in the returned read. When those check out, it is independent evidence that the read came from the reaction you think it did.
Yes. The full report downloads as a plain text file carrying the alignment, the identity and coverage numbers, the probability of matching by chance, the primer evidence, the verdict and a written justification of each conclusion. Saving it at the time the result arrives keeps the evidence with the sample instead of reconstructing it weeks later.
All processing happens locally in your browser. Nothing you upload is transmitted or stored, as described in the privacy policy.