I have two main modeling research lines: one related to structural biology (including rational drug design, de novo design, polymorphism, etc.) and the other one related to condensed matter (nanostructures, crystals, etc.).

Any time I submit a manuscript in the biological area, either the editors or the reviewers ask for experimental validation, even when we use very well established simulation tools to make the predictions (here we work only with small molecules and proteins, not with living organism simulations). On the other hand, when submitting a manuscript in the condensed matter area, we never received that complaint.

Why are condensed matter predictions considered on par with experiment, while structural biology modeling receives more skepticism?

  • $\begingroup$ The first two paragraphs (and title) are about a comparison between structural biology and condensed matter, but the last paragraph (in bold, and actually with a question mark) asks for examples where matter modeling successfully predicted an experimental result. So what are you looking for here? Dirac's equation predicted the positron in 1928, and this turned out to be a successful prediction of its 1929 observation by Dmitri Skobeltsyn. I could turn that into an answer, and we can maybe have a list of dozens of examples from other people. It would be a good resource: couldn't find on google. $\endgroup$ Jul 29, 2020 at 2:26
  • 2
    $\begingroup$ @NikeDattani I think the bolded question is just to give a more concrete direction for an answer, but isn't necessarily required. I think the crux of the question is something like "why are condensed matter predictions considered on par with experiment, while structural biology modeling receives more skepticism?" $\endgroup$
    – Tyberius
    Jul 29, 2020 at 21:03
  • $\begingroup$ You nailed it @Tyberius. Please, feel free to edit the question. $\endgroup$
    – Camps
    Jul 29, 2020 at 22:58
  • 2
    $\begingroup$ I don't think drug discovery can be equated with structural biology. I'm curious to know for what properties they demand you obtain experimental data in the case of structural biology. $\endgroup$
    – Buck Thorn
    Jul 30, 2020 at 18:09
  • $\begingroup$ What is the opposite of rational design? and so it goes... catch phrases like rational design are just hand waves. Nobody is doing irrational designs... when people say they are doing rational designs... they are just using a buzz word for hand waving $\endgroup$
    – B. Kelly
    Dec 20, 2021 at 3:25

1 Answer 1


If this is happening, I think there's a few reasons:

Quantum physics in general is the science subject where the most precise agreement between theory and experiment exists. For example, the experimental work: Nature 588, pg 61–65 (2020) "Determination of the fine-structure constant with an accuracy of 81 parts per trillion", and a QED calculation involving 12672 10th order Feynman diagrams agree in the "part per 10 billion" digit for the reciprocal of the fine structure constant :

1/137.0359992 [experiment, up to 10 significant digits]
1/137.0359992 [QED, up to 10 significant digits]

Even for actual matter modeling, we have had similar success, see: How accurate are the most accurate calculations?. Solid-state condensed matter is more complicated since more atoms and molecules are involved, but we are still able to quite precisely predict phase transitions, conductivity, and with BCS theory, even superconductivity properties, often within the limits of what experimentalists can do, or to what practically matters. The editors of physics journals, and the people from the physics community that will be asked to referee papers, will largely come from physics departments, which at least until very recently had far more people from the "old-school" physics tradition ("biophysics" is a relatively new term) and will be used to trusting calculations.

Biology is on average much less quantitative. Some areas can be extremely quantitative, but the vast majority of a biology department will be wet lab experimentalists who did not learn math or computer programming beyond their first year calculus class, rather than bioinformatics (another relatively new term) researchers. Most biologists are used to being around people who don't have the same level of appreciation for the fact that things like "$\hat{H} | \psi_n(t) \rangle = i \hbar \frac{\partial}{\partial t} | \psi_n(t) \rangle$" can sometimes actually predict physical quantities with more accuracy than the best experiments (something that is often proven to be correct many years or decades later).

Biology by definition is about living things, and life is a complicated thing (even the components of life, which you're studying in a structural biology simulation). It's far easier thermodynamically, for atoms and molecules to randomly come together and form a solid-state material that can be studied by typical condensed matter physicists, than for randomness to result in a non-living protein that can be studied by typical structural biologists. In that sense, biologists are a lot less lucky, in that they are not dealing with quantities for which we can calculate the values ab initio with such exquisite accuracy: I often say "climate change (like biology) is complicated because we don't have the Schroedinger equation for tomorrow's weather" and I also happened to see this quote at the end of a youtube video this morning:

"I can calculate the motion of heavenly bodies but not the madness of people." ~ Isaac Newton

I can also calculate the conductivity of a material, but not how a protein folds.

  • 1
    $\begingroup$ I think the example of AlphaFold and the subsequent community modification and replication (RosettaTTAFold, etc.) is a good example. People are now routinely using these methods for interesting bioscience. $\endgroup$ Dec 21, 2021 at 19:31

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .