Calculator Inputs
Plotly Graph
The chart compares bit counts, union size, and the calculated similarity level for the current submission.
Example Data Table
| Example Pair | Bits in A | Bits in B | Shared Bits | Union Bits | Tanimoto Coefficient | Similarity % |
|---|---|---|---|---|---|---|
| Fingerprint Example 1 | 128 | 140 | 96 | 172 | 0.5581 | 55.81% |
| Fingerprint Example 2 | 150 | 150 | 120 | 180 | 0.6667 | 66.67% |
| Fingerprint Example 3 | 200 | 180 | 150 | 230 | 0.6522 | 65.22% |
Formula Used
T = c / (a + b - c)
D = 1 - T
a = active bits in Compound A, b = active bits in Compound B, c = shared active bits, and (a + b - c) = union bits.
In binary chemistry fingerprints, the Tanimoto coefficient measures the overlap between two structures relative to all active features observed across both structures.
How to Use This Calculator
- Enter a name for each compound or fingerprint record.
- Provide the active bit count for Compound A and Compound B.
- Enter the number of shared active bits observed in both fingerprints.
- Set the fingerprint length, the comparison threshold, and desired decimal precision.
- Add optional descriptors and report notes for clearer exported output.
- Click Calculate Similarity to display the result above the form.
- Review the table, metrics, and graph, then export the report as CSV or PDF.
Frequently Asked Questions
1) What does the Tanimoto coefficient measure?
It measures how strongly two molecular fingerprints overlap. A higher value means more shared structural features relative to the total active features across both fingerprints.
2) Is Tanimoto similarity the same as Jaccard similarity?
For binary fingerprints, yes. In chemistry workflows, Tanimoto is commonly used as the practical form of Jaccard similarity for bit-based structure comparisons.
3) Why must shared bits be less than both bit counts?
Shared active features cannot exceed the number of active features present in either compound. If they do, the input combination is physically inconsistent.
4) What is a good similarity threshold?
Thresholds depend on your screening purpose and fingerprint design. Many teams test several cutoffs, then pick the value that best matches their retrieval or clustering goals.
5) What is the difference between Tanimoto and Dice?
Both compare overlap, but they scale it differently. Tanimoto uses the union size, while Dice doubles the shared count and divides by the combined active counts.
6) Can I use decimal values instead of integers?
Yes. This page accepts decimal summaries, which can help when you are using averaged counts, weighted fingerprints, or normalized feature totals.
7) Why is fingerprint length included?
Fingerprint length provides occupancy and coverage context. It helps you judge how dense each fingerprint is relative to the available descriptor space.
8) What does the classification label mean?
The label compares the calculated coefficient with your chosen threshold. It quickly shows whether the pair is above, near, or below the selected target.