Data
Heat ladder explorer
Five orders of magnitude of engineered heat, laid out so you can see where the catalog is dense, where it is thin, and which rungs rest on measurement rather than prediction.
100 – 16M
SHU span
6
tiers
70
series
1,000
constructs
01
What the dataset is
A thousand constructs placed on a logarithmic axis running from 100 SHU to 16,000,000 SHU. Six tiers divide it — Pepper at the bottom, then Tier 1 through Tier 5 — and seventy series cut across the tiers by design intent rather than by heat. A series is a family with a shared approach; a tier is a band of intensity. Most useful questions involve both.
The axis has to be logarithmic, because the ladder spans five orders of magnitude. A two-hundred-SHU difference is the whole identity of a design at the bottom and invisible rounding at the top. The explorer keeps that asymmetry in front of you rather than flattening it.
02
How to read it
Measured versus predicted
Entries backed by an HPLC heat rating on grown material are marked distinctly from entries carrying a model prediction. This distinction is never collapsed, because it is the one that matters most.
Density shows maturity
Crowded regions of the ladder are well-characterized. Sparse regions are where extrapolation starts, and the explorer makes that visible before you design into one.
Host colour
Entries carry their host organism, so you can see at a glance which parts of the ladder are pepper-only and which have been reached in cross-species hosts.
Burn profile, not just total
Each entry carries its capsaicinoid ratio and the burn profile that follows from it. Two entries at the same rung can be entirely different experiences, and the profile column is where that shows.
Series as a thread
Filtering to one of the 70 series pulls out a design lineage across tiers — the same approach taken progressively further, which is often more informative than a cross-section.
Neighbours on demand
Select any point and the explorer returns its nearest neighbours by heat, by host and by ratio. This is the fastest way to find out whether a new target has precedent.
03
A worked look
- View
- Tier 2 band, cross-species hosts
- Heat window
- 1,000 – 10,000 SHU
- Hosts shown
- Tomato, tobacco, 9 others
- Entries
- Including SS-0421
- Ratings
- Mixed measured and predicted
- Sort
- By burn profile, then host
- Export
- CSV with provenance column
Tier 2 in cross-species hosts is a good band to study because it is where the catalog’s cross-species work is most developed. The entries cluster, and the clustering is not accidental — it reflects where host flux budgets comfortably land once the pathway is installed. Looking at this band before setting a tomato target tells you more than any single prediction will.
The provenance column is the one to keep an eye on. A dense cluster of predictions is a different kind of evidence from a sparse set of measurements, and the explorer refuses to let the two blur together.
04
What you can learn from it
Where on the ladder a new target has precedent and where it is genuinely novel. Which hosts have been taken how far. How burn profile varies independently of total heat. Which rungs rest on HPLC measurement and which are still predictions waiting for material.
The explorer is free to browse and free to export. It is deliberately the first thing most projects open, because the cheapest way to improve a target is to look at what already exists near it.
Ready to run it
Try this in the Splice Platform.
CPU engines are free to run. GPU engines run on prepaid credits.
Related