Essays and findings for cardiac-safety scientists. Each piece makes one argument: where a single-channel margin fails, why mechanism matters, what a mechanism read returns, and the validation behind it. The failures are published as often as the wins.
Astemizole carried a 128-fold hERG margin and was pulled from the market for TdP. The margin reads it Low; the model reads it Intermediate; the market read it as dangerous. The whole case for mechanism on one page.
Verapamil blocks hERG potently and never causes TdP in the clinic, because it blocks calcium at the same time. hERG-only screening calls it High. The mechanism read calls it Intermediate. The calcium limitation, confirmed in three datasets, is on the page.
The pipeline, the two halves, and the one experiment that confirms the cause. What you receive from a read, and the scope of each part.
On the 28-drug CiPA panel, a plain hERG margin gets 18 of 28 right and we get 17. The loss is on the site, with the compounds where each method is wrong. This is the trust anchor: a tool that publishes its own failure.
The bioRxiv preprint, in full. Against Blinova et al. 2018, 112 matched drug-dose points give a rank correlation of 0.67 and direction agreement on 89% of measurable points. Against Lee et al. 2025, direction agrees on 9 of 10 anchored drugs.
The first read on a public or non-confidential compound is free. Same report format as the paid read, so you can judge the work before you buy.