DNA fingerprint

This is a sequence-composition image, not a forensic DNA profile or proof of identity. Paste DNA/RNA or one or more FASTA records, choose k (1–9), then build. A cell counts one overlapping word of exactly k bases. Records are counted separately and their counts are added; no word crosses a FASTA boundary.

Orientation is standard CGR: C/G at the top, A/T at the bottom. The last base chooses the largest quadrant; earlier bases refine the position. Click a cell or type a k-mer and Inspect. Count and frequency always refer to exact k-mers, with frequency divided by the number of valid windows. The outline marks the selected cell; PNG exports the raw matrix without this overlay or corner labels.

IUPAC ambiguous bases R Y S W K M B D H V N are accepted as separators: every window containing one is skipped, never guessed or joined across. RNA U is converted to T; mixed T/U within a record, numbers, gaps and empty records are rejected. Counts use the entered strand only; reverse complements are not merged.

Black means zero, white the largest value in this image. Logarithmic intensity uses log(1+count) to reveal less frequent words; it does not change counts or CSV. In Options, summing lengths 1…k overlays each smaller-length count on its matching cells. This is a hierarchy score, not a k-mer frequency. The image is normalized separately, so brightness alone cannot compare different sequences.

PNG and image copy retain the 2ᵏ×2ᵏ raster. CSV exports observed k-mers only, with count and relative frequency; omitted words have count zero. The GC-rich example is deterministic synthetic data, not an organism. The initial ACGT repetition is also synthetic.

Local worker processing, with real counting progress and cancellation. Limit: 2 million bases and 3 million input characters. Editing sequence, k or hierarchy removes stale results and stops the worker. Intensity changes redraw the same counts. No permanent storage or network upload.