SNP / Indel Analysis · tools.omics.bd

1 Sequences FASTA or one per line · max 100,000 bp each

Paste sequences

Provide at least two sequences. Headers (>) are optional — without them each non-empty line is one sequence. Both DNA (ACGT) and RNA (ACGU) are accepted; IUPAC ambiguity codes (R, Y, S, W, K, M, B, D, H, V, N) are supported.

Analysis options

Reference sequence

All other sequences are aligned to this one.

Anchor k-mer size

Larger = more specific, may fail on divergent sequences.

Min allele frequency (%)

Hide variants present in fewer than this % of samples.

Report

Full per-position tables can be very large.

i How the analysis works

Alignment — a seed-and-extend strategy: unique k-mers (default 15 bp) act as anchors; gaps between anchors are filled with Needleman–Wunsch.
SNP — single-base substitution where reference and query are not IUPAC-compatible.
Insertion — bases present in the query but absent from the reference.
Deletion — bases present in the reference but absent from the query.
Consensus — for each variant site the union of observed bases is collapsed to the appropriate IUPAC symbol.
Filtering — the minimum allele frequency hides variants below the chosen fraction of samples.
Limits — maximum 100,000 bp per sequence.

📘 Tutorial & Result Interpretation

📑 On this page

What this tool does — 2-minute overview
IUPAC ambiguity codes — the reference table
How the alignment works
Preparing your input — best practices
Reading the variant table
Reading the IUPAC consensus
Reading the pairwise difference matrix
Using the alignment viewer
Recommended workflow — Sanger amplicon analysis
Common pitfalls & best practices

🧬 1. What this tool does

This tool aligns two or more DNA sequences against a reference, then reports every position at which the samples differ from the reference — and from each other. It is designed for Sanger amplicon analysis: a typical use case is resequencing a target gene in several individuals and asking which positions vary.

SNP — Single Nucleotide Polymorphism: one base differs.
Insertion (INS) — extra bases in a sample that are absent from the reference.
Deletion (DEL) — bases in the reference that a sample lacks.
Consensus — one IUPAC-encoded sequence that collapses all samples into a single string.

Key idea: the tool does not decide what is "real". It reports what it sees. Interpretation — whether a variation is a true biological variant or a sequencing artefact — is your job.

🔤 2. IUPAC ambiguity codes

When different samples carry different bases at the same position, the consensus cannot be a single letter. IUPAC ambiguity codes let you express "any of several bases" in one character.

A A · Adenine

C C · Cytosine

G G · Guanine

T T · Thymine

R R · A or G (purine)

Y Y · C or T (pyrimidine)

S S · G or C (strong)

W W · A or T (weak)

K K · G or T (keto)

M M · A or C (amino)

B B · C, G or T (not A)

D D · A, G or T (not C)

H H · A, C or T (not G)

V V · A, C or G (not T)

N N · any base

You see	It means	Typical cause
A, C, G, T	All samples agree on this base.	Invariant position.
R	A and G both observed.	Transition A↔G SNP.
Y	C and T both observed.	Transition C↔T SNP.
S	G and C both observed.	Transversion G↔C SNP.
W	A and T both observed.	Transversion A↔T SNP.
K	G and T both observed.	Transversion G↔T SNP.
M	A and C both observed.	Transversion A↔C SNP.
N	Any base — cannot decide.	Poor quality, or ≥ 3 alleles.

In diploid organisms, an R at a position often means the individual is heterozygous for an A/G SNP.

📐 3. How the alignment works

Alignment is the process of inserting gaps into sequences so that homologous positions line up. This tool uses two strategies:

Identical length — position-by-position

If a query is the same length as the reference, it is compared base by base. Fast and exact.

Different length — anchor & fill

Unique k-mers (default 15 bp) are indexed in the reference.
The same k-mers are searched in the query — those matches are anchors.
Anchors are chained left-to-right to form a skeleton.
The gaps between anchors are filled by Needleman–Wunsch pairwise alignment.

This is much faster than full NW for large inputs and produces the same result when sequences are highly similar (> 90 % identity), which is the normal case for Sanger amplicons.

k-mer size matters. Smaller k = more anchors but more false matches; larger k = fewer anchors, may fail on divergent sequences. Default 15 is good for sequences ≥ 95 % identical. If alignment fails, try k = 12.

📥 4. Preparing your input

FASTA format (recommended)

Standard FASTA — one header line per sequence, then the sequence on as many lines as needed:

>sample1_forward
ATGCGTACGTAGCTAGCTAGCTAGCTAGCATCGATCGTAGCTAGCTAGCTAG
CATCGATCGATCGATCGATCGATCG

>sample2_forward
ATGCGTACGTAGCTAGCTAGCTAGCTAGCATCGATCGTAGCTAGCTAGCTAG
CATCGATCGATCGATCGATCGATCG

Plain one-per-line format

Each non-empty line is treated as one sequence, named Seq1, Seq2…

What this tool accepts

DNA — ACGT.
RNA — ACGU; U is converted to T.
IUPAC ambiguity codes — R, Y, S, W, K, M, B, D, H, V, N.
Anything else is stripped. Whitespace, digits, and punctuation are ignored.

Forward + reverse reads? Reverse-complement them yourself before pasting, so all sequences are on the same strand. The tool does not auto-orient.

📊 5. Reading the variant table

Column	Meaning
Position	1-based coordinate in the reference.
Type	SNP / Insertion / Deletion; sub-label shows the length in bp.
Reference	The base(s) present in the reference at this position.
Allele(s) observed	Every alternative allele seen across the samples, with a count chip for each.
Consensus	IUPAC code representing all observed alleles at this position.
Samples affected	Number of samples (and their names) that carry any non-reference allele here.

How to decide whether a variant is "real"

Present in multiple samples? A variant that appears in ≥ 2 independent samples is more convincing than a singlet.
In the same position as a known variant? Cross-reference with dbSNP if you are working with a well-characterised organism.
Supported by both strands? A heterozygous C/T SNP is visible on both forward and reverse reads. An artefact is often only on one strand.
Present at the very start or end of the read? Poor quality; usually discarded.

Mono-allelic indel artefact. A single deletion in one sample at a homopolymer run is almost always a sequencing artefact. Raise the minimum allele frequency filter to remove singletons.

🧾 6. The IUPAC consensus

The consensus is a single sequence that represents all samples. At invariant positions it matches the reference; at variant positions it uses an IUPAC ambiguity code combining all observed bases.

Example

Reference:  ATGCTAGCTAGC
Sample 1:   ATGCTAGCTGGC
Sample 2:   ATGTTAGCTAGC
            ────────────
Consensus:  ATGYTAGCTRGC

Position 4 — C in ref, C in S1, T in S2 → Y (C or T).
Position 10 — A in ref, G in S1, A in S2 → R (A or G).

When to use it

Primer design in a variable region — pick a conserved region, but keep track of ambiguity.
Submitting to GenBank — for population samples the IUPAC string is a compact representation.
Reference for the next round of analysis — paste it back as the reference in a follow-up run.

Never submit an IUPAC consensus as if it were a single sample's sequence. It is a representation, not an observation.

📋 7. Pairwise difference matrix

The matrix shows, for every pair of samples, the percentage of variant positions at which they differ (relative to the reference length).

Value	Interpretation
< 1 %	Nearly identical — same haplotype or very closely related.
1–5 %	Moderately different — different haplotypes, or different species of the same genus.
> 5 %	Substantially different — different species, or a misalignment.

Quick sanity check: the diagonal must always be 0. If two samples are identical, their off-diagonal cell should be 0 too — if it isn't, the alignment may have artefacts.

🔍 8. Alignment viewer

The viewer shows the pairwise alignment of one sample against the reference, one block per variant site. Varying positions are boxed and coloured by type.

Reference row — the reference sequence. · means the reference has a gap here.
Sample row — the aligned query. · means the sample has a gap here.
Match row — | for a match, • for a mismatch.
Colours — red = SNP, green = insertion, amber = deletion.

Zoom in by reducing the window size to 10–15 bp to see the exact local sequence around each variant.

🚀 9. Recommended workflow (Sanger amplicon)

Sequence your PCR products on both strands.
Trim low-quality ends with a base-caller or a tool like SeqTrace.
Reverse-complement reverse reads so all sequences are on the same strand.
Paste all samples (FASTA format) into the input box.
Set the reference — usually the wild-type or the annotated reference sequence.
Run. Inspect the variant table; note the positions and the sample names.
Filter: if the sample set is small, raise the minimum allele frequency to remove singletons that are likely sequencing errors.
Download the consensus if you need a single representative sequence.
Cross-check — look at the chromatograms for every reported variant.
Confirm by an orthogonal method: RFLP, SNaPshot, HRM, or NGS.

⚠️ 10. Common pitfalls

Reverse read not complemented. The most common mistake. Forward and reverse reads must be on the same strand. If you paste the reverse read as-is, every position looks like a variant.

Chromatogram artefacts at the ends. The first 20–30 bp and last 20–30 bp of a Sanger read are unreliable. Trim them before analysis.

Heterozygous indels. A single-base deletion in a homopolymer run is one of the hardest features to call from Sanger data. Expect an IUPAC consensus with ambiguous calls nearby.

Low coverage positions. If a variant appears only once across many samples, verify the chromatogram — it may be a chemistry artefact.

Recombination of paralogues. If two amplicons are actually products of two related genes, alignment will show scattered variation across the whole length. Re-check primer specificity.

Best practice: sequence both strands, trim ends, complement reverse reads, and treat any singlet variant with scepticism until the chromatogram is inspected.

QC checklist

✔ All sequences on the same strand (reverse reads complemented)
✔ Low-quality ends trimmed
✔ ≥ 2 independent samples per reported variant (or chromatogram verified)
✔ Reference sequence is correct and annotated
✔ Matrix diagonal = 0 (sanity)
✔ Consensus exported with a clear name — e.g. genename_population_consensus.fa
✔ Variants confirmed by an orthogonal method before publication