A Field of Research · Anno MMXXVI

Tunnelling Informatics

The data-driven study of hydrogen tunnelling in enzymes
Sabrina Hackfort · Independent Researcher · ORCID 0009-0007-7458-2487
❦ ❦ ❦

Not to be confused with the information systems of tunnel construction, nor with network tunnelling in computer science, such as VPN or SSH tunnels. The tunnel here is the quantum-mechanical one.

Enzymes move hydrogen — as a proton, as a hydride ion, as a hydrogen atom — and part of it often does not climb over the energy barrier of the reaction but passes through it. Since 1989, when Cha, Murray and Klinman reported isotope effects in yeast alcohol dehydrogenase that pointed to tunnelling, this has been read mainly from kinetic isotope effects: how much slower a reaction runs when deuterium or tritium takes the place of hydrogen, and how that ratio changes with temperature. These numbers are scattered over hundreds of papers, written in different conventions, and often cited on without their conditions. Tunnelling informatics treats them as data.

Definition

Tunnelling informatics (also tunneling informatics) is the data-driven study of hydrogen tunnelling in enzymes. It collects the measured kinetic isotope effects with their sources and conditions, establishes what such data can and cannot prove about tunnelling, re-examines the published claims against them, and tests predictions made from enzyme structure. Every conclusion is to follow from open data and from rules stated in advance.

The name was introduced by Sabrina Hackfort in 2026.

Four Parts

I Data  ·  II Proof  ·  III Re-examination  ·  IV Prediction
  1. The data. MEG-NEN, a curated database of the measured kinetic isotope effects of enzymes: every value with its source, isotopes, temperature, pH and conventions, and every difference in activation energy brought to one sign convention. Each value is read twice, the second reading blind to the first, before it is accepted; the five that only one reading found say so in the row. Values that a paper only cites from other work, or shows only as a point on a plot, or only in its supporting information, are left out — a number printed in a figure legend is printed text and is kept.
  2. What the data can prove. An identifiability result: which statements about tunnelling such measurements can establish, and which no number of them can. An isotope effect is no direct measure of tunnelling, since zero-point energy alone already makes hydrogen react faster than deuterium (see ZENO).
  3. The re-examination. The claims made since 1989, checked against the data: where the measurements support them, where they contradict them, and where they cannot decide.
  4. The prediction. Predictions from an enzyme's structure, made before the measurement and then confirmed or refuted in the laboratory. A first structure-based classifier, MEG-APSU, exists; its honest accuracy and its limits are on its page.

Where It Stands

On 10 November 2026, MEG-NEN held 297 kinetic isotope effects and 167 temperature series from 62 papers, all but five values read twice; four further papers on the reading list have not yet been obtained. It is under construction. The working repository has been public since 11 November 2026, and a data note describes how the database is built, how far it has been checked, and what has not yet been checked. A first check of a sample of 47 rows consisted of a third machine reading and a comparison with the sources by the author, which was not blind and is recorded row by row for 11 of the 47. A numbered release with its own DOI will follow once the outstanding papers have been obtained and the open items listed in the note have been closed. The identifiability result, the re-examination and the prediction build on it, in that order.

Related Work

How to Cite

Hackfort, S. (2026). Tunnelling informatics: the data-driven study of hydrogen tunnelling in enzymes. https://sectio-aurea-q.github.io/tunnelling-informatics.html

Reference: Y. Cha, C. J. Murray, J. P. Klinman, Hydrogen tunneling in enzyme reactions, Science 243, 1325–1330 (1989), doi:10.1126/science.2646716.