Find a neighbourhood
Computed in your browserLoading the local snapshot and building the search index…
Sequence neighbourhood
Select a node to inspect its recordEdge thickness reflects similarity. Node positions only arrange the results; they are not an embedding.
Nearest sequences
One row per normalized sequence. Open a row to see all matching source records.
| # | Similarity | Sequence / target | Target type | Affinity as reported | Records |
|---|
Shared sequence motif
6-mer anchor + flanking basesFirst occurrence of the most-shared query 6-mer, with six bases on each side. Only sequences containing that motif contribute. This is an exploratory alignment; shared primers and repeated bases can dominate.
Record details
Browse the full collection · includes annotated and ambiguous sequences
| Record | Sequence ID | Target | Year | Search eligibility |
|---|
How to read these results
- Sequence similarity, not a binding prediction
- Targets come from the source records. Sequence resemblance and target frequency do not establish binding or affinity for your query.
- Exact edit-distance
- 1 − Levenshtein distance / the longer sequence length. Insertions, deletions, and substitutions each cost one. Scores range from 0 to 1.
- k-mer cosine
- Frequencies of all 1–4-mers (340 features), standardized using the unique searchable sequences in this snapshot, then compared by cosine similarity. Scores range from −1 to 1. This captures composition and is not percent sequence identity.
- Sequence handling
- Case and whitespace are normalized, and U is mapped to T for scoring and grouping. Other symbols, ambiguous bases, and modification annotations are excluded from scoring and retained in the full collection browser. Exact-match hiding compares normalized sequences, including when using k-mer cosine.
- Records and filters
- Duplicate normalized sequences share one neighbour. Filters select source records before ranking; details retain their provenance. Alphabet filters describe letters present in the sequence, not confirmed DNA/RNA chemistry. Target types and level codes are displayed as supplied, without inferred labels.
- Affinity and structures
- Affinity text, pKd, MFE, and dot-bracket structures are shown as supplied. Structure diagrams are the downloaded source images; this viewer does not predict new structures. Source annotations have not been independently validated.