I'm trying to model the reaction mechanism of an enzymatic steroid reduction. Experimental research proposes a large variety of intermediates, which I believe would each require separate parametrization through AMBER.

Using a PM7 medium-layer, I should be able to achieve a near-linear cost modelling of non-covalent interactions between the SUB/COF.

The reason for using AMBER over UFF is mostly those non-covalent interactions, from my understanding.

PM7, while not the best of levels of theory, should still be better than MM for getting accurate results from what papers I found.

The M06-2X functional is picked due to its performance in both non-covalent interactions and transition state energy calculations.

To note, the only paper that did a comparable study to mine was published 2021 May. Before that, people used (B3LYP/6-31G*:OPLS) for fairly low-accuracy results.

Would potential reviewers in the field take issue to using a deviation from existing methods if my results align with experimental data?

  • 1
    $\begingroup$ Is the meat of your question the final sentence? If it is, the answer is that you won't know until you submit the publication. The referees will/should be knowledgeable in the field and will offer their opinions. $\endgroup$ Commented Mar 29, 2022 at 16:11
  • $\begingroup$ It is the meat of it, yes. I'm trying to determine whether to invest in doing all the AMBER parametization, or focus on the already involved computation. $\endgroup$
    – Runa
    Commented Mar 30, 2022 at 2:14
  • 3
    $\begingroup$ At the risk of being biased (I am an AMBER disciple), in general if there is an accepted standard you think the referees will lean towards (and your new method yields superior results), it can't hurt to front-run that and do the comparison now. It will probably come up at some point anyway and best to get it out of the way and establish your new benchmark. $\endgroup$ Commented Mar 30, 2022 at 2:49
  • 1
    $\begingroup$ I was working in that field and I am still not accustomed to all of the used abbreviations here. Better explain or introduce them. I'm, however, going with my gut here: Anything is better than B3LYP. Quite importantly will be justification and calibration; if you want to trailblaze, then start doing it right. (And also have a look at xtb, please.) || Anyway, this might have been a better fit at Matter Modeling and future questions of this kind might get better attention/answers there. It's is a good resource in any case. $\endgroup$ Commented Mar 30, 2022 at 21:49


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.