I know it is probably a case by case basis, but how low of an "Accepted Energy" reading would be considered LOW?

I see a section for the lowest energy structure predicted, and was wondering how low of a reading anyone has seen on a particuar WU.

As I am not very good at searching through everyone's old posts, if there was an old post like this one, could someone provide a link, as I would like to read about it.

---

The predictor of the day (as published on the home page in the news) is basically the lowest energy level found for all of us on a given work unit. Since the current proteins are CASP, I believe they aren't publishing exactly what the energy level of that best prediction was. I'm sure they will publish more once CASP is over.

I've got one with -425. The longer the protein is, the more force it takes to keep the whole thing from falling apart. Least that's how I think of it. For a shorter protein, -75 might be the energy level of the native form.

Just based on my observation of the WUs crunched, it seems as though -1.5 times the number of amino acids in the sequence seems to be about as low as you see.

Also, after CASP, they will work on bringing online a graphing system that will show all of our predictions and highlight those made by you, and those made by your team. Or... maybe not ALL of your predictions, but at least your BEST one. It's pretty neat, but the initial coding done to deliver it wasn't efficient, and the resulting server load was too great. So they have to devise a better approach to making the graph and plotting your results on it.

Here is an example of what the graphs look like. See that little blue X in the lower left? That's the lowest energy, or RMSD. The red + signs are predictions made by other participants using a different random starting point on the protein.

I also wanted to point out that once the scale on your low energy box reaches your prior predictions, they appear there with a red box. Those are YOUR actual predictions reached on model 1 and 2 and 3 that you are seeing there while you are crunching model 4.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

Wait a sec..

Low energy and RMSD are different. The algorithm we're using tests our models/decoys/predictions and gives an energy value. The lower the energy, the better chance it's a candidate for low RMSD. (Although we've had graphs where the lowest RMSD was 2/3rds up the graph, instead of on the absolute bottom with the lowest energy models/decoys/predictions.)

The RMSD readings we've been seeing were RMSD from the native structure. We can't determine RMSD now, because we'd need a native structure to compute the root-mean-square deviation from the native structure. i.e. How close is each amino acid to the correct placement on the native structure.

---

We can't determine RMSD now, because we'd need a native structure to compute the root-mean-square deviation from the native structure.

Exactly. I was just trying to explain what you'll see when you follow the links. There is a graph for "lowest energy", and another for "lowest RSMD". The protein being displayed there has a known structure, and so the models produced were compared to that known structure as well as showing the energy level of the model produced.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/