Add beta_jan25 energy function

[Haddox]

Hi @roccomoretti and @lyskov,

@fdimaio and I have developed a new beta energy function called beta_jan25. This is an updated version of beta_nov16.

The aim of this PR is to add beta_jan25 to the rosetta source code.

We will shortly post a manuscript describing how we developed beta_jan25. They key updates are to the LJ potential. We identified steric clashing in proteins that were relaxed or designed using beta_nov16. We identified examples of this problem in a high-quality benchmark from the dualoptE protocol used to train the energy function. We then used this benchmark, and the others in dualoptE, to refit a small number of LJ parameters. The refitting largely eliminated the clashing problem, and beta_jan25 is as good or better than beta_nov16 when assessed on multiple benchmarks using validation data.

Please let me know if you have any comments or questions. I am also happy to share a draft of the manuscript. I am having Frank review my changes, and then I can ping one or both of you when Frank gives the thumbs up.

@fdimaio: could you please review my changes? And let me know if there are any additional changes I need to make? A few questions I had are:

• should genpot be "on top" of beta_jan25? Or should we stick with beta_nov16? See line 752 of score_function_corrections.cc
• currently, the -beta flag does not give the same results as either the -beta_jan25 flag or the -beta_nov16 flag. At least one reason for this is it seems to invoke the beta_genpot.wts file. Is this the correct behavior?

Thanks!

Hugh

[roccomoretti]

Looks reasonable from my end. -- My one recommendation would be to add tests to make sure -beta_jan25 is invokable properly. (You can take a look at the unit tests in source/test/ and the integration tests in tests/integration/tests/ to see if any of the tests using -beta_nov16, -beta or -gen_potential can be copied for -beta_jan25).

I'll leave it to Frank to decide if -gen_potential should use beta_nov16 or beta_jan25 (and what plain -beta should invoke.)

[roccomoretti]

Please update the comments at the top to be accurate

[Haddox]

Good catch! Just pushed updates.

[lyskov]

Looks good to me! @Haddox please let us know when this is ready to be merged. Thanks,

[Haddox]

Thanks @lyskov!

@roccomoretti: Thanks for the suggestion to add tests to make sure beta_jan25 is invoked properly. After discussing it with Frank, we decided to add a test in tests/sfxn_fingerprint/scores/, which scores a set of input PDBs with a specific energy function and uses the output scores as a fingerprint. The test makes sure this fingerprint remains the same over time. I just added a version of this test for beta_jan25 in my last commit.

Frank also advised me that -gen_potential should stay the same (on top of beta_nov16), and we decided to have plain -beta invoke beta_jan25.

Updating -beta will cause some of the previous tests to fail because it now calls a new score function. Frank suggested I run the tests, and then Frank and I will look over the results to make sure any failing tests look reasonable (small changes in numbers due to the new score function). I will do that now.