Rosetta 3.13 seems very interesting:

- Scientific benchmarking system rejuvenated (and submitted for publication)
- trRosetta available in C++ Rosetta now. A TensorFlow build (extras=tensorflow or extras=tensorflow_gpu) supports this
- Support for new Mac M1 chipset
- RosettaSurf (protein surfaces)
- PyMOL-Rosetta hookup can now be "bounced" off a Gray lab server to allow viewing of Rosetta trajectories when the trajectory source is insufficiently configurable
etc

Waiting for new science in Rosetta@Home!

---

extras=tensorflow_gpu

I am sure that caught your eye.

---

extras=tensorflow_gpu

I am sure that caught your eye.

:-P
I think they will use this option internaly, to "train" neural networks.
Despite all, seems a big step for the project

---

I think they will use this option internaly, to "train" neural networks.

After "The DeepMind tornado" in CASP14, seems that AI/ML/Neural networks are the new "holy grail" in protein's field, ex like this

---

It's a shame, that there is no communication and announcement from the project side regarding the new application and state. For me it's no appreciation as cruncher to get no info regarding the state, and also make the project very less atractive. For me, it looks like a blind flight project.

I have lost some team mates regarding the non fixed problems with the Rosetta applicaitons, which caused some hung of there PCs. So they deinstalled BOINC.

Due this, I have decided to reduce the resource share for Rosetta and Ralph and gave the computing power to some more valuable projects.

---

Due this, I have decided to reduce the resource share for Rosetta and Ralph and gave the computing power to some more valuable projects.

I am in the same boat, especially during the summer when I have to turn off some machines and readjust the work. Rosetta will get less, but still some.
I strongly suspect that they will go the way of WCG/MIP (since they are based on Rosetta), and cut back, if not eliminate, the BOINC work and do more inhouse.
https://www.worldcommunitygrid.org/forums/wcg/viewthread_thread,43484

But they didn't bother to tell us even about the new Python app. So we will find out when we find out.

---

I strongly suspect that they will go the way of WCG/MIP (since they are based on Rosetta), and cut back, if not eliminate, the BOINC work and do more inhouse.
But they didn't bother to tell us even about the new Python app. So we will find out when we find out.

I don't think they eliminate boinc work (at least soon), otherwise why create the new Python app?
By the way, the lack of communication is frustrating

---

I don't think they eliminate boinc work (at least soon), otherwise why create the new Python app?

If the amount of Python generated thus far is any indication, it is very much a minority project.
Maybe it is used for training the AI stuff?

I would be happy to do it, but I doubt we would get much.

---

If the amount of Python generated thus far is any indication, it is very much a minority project.

Or, maybe, is just a beginning and researchers are still trying to understand new app and results.
But, as i continue to repeat, without info we are "blinds"

---

Interesting news about Rosetta code also on Foldit:

Hey everyone!

We're about to push an update to the developer preview with some big changes to our underlying biochemistry software, Rosetta.

Rosetta provides Foldit with algorithms that represent the state-of-the-art of protein biochemistry modeling. However, the version of Rosetta we are using is now quite a few years out of date, and we'd like to get back to the cutting edge.

This update brings us up-to-date with the latest version of Rosetta. You hopefully shouldn't actually notice any changes... yet. The benefits of this update will come down the road as we begin to take advantage of the features in this more capable version of Rosetta.

If you do notice any behavioral differences in your testing, please leave us some feedback!

---

Interesting news about Rosetta code also on Foldit:

Hey everyone!

We're about to push an update to the developer preview with some big changes to our underlying biochemistry software, Rosetta.

Rosetta provides Foldit with algorithms that represent the state-of-the-art of protein biochemistry modeling. However, the version of Rosetta we are using is now quite a few years out of date, and we'd like to get back to the cutting edge.

This update brings us up-to-date with the latest version of Rosetta. You hopefully shouldn't actually notice any changes... yet. The benefits of this update will come down the road as we begin to take advantage of the features in this more capable version of Rosetta.

If you do notice any behavioral differences in your testing, please leave us some feedback!

Cool!!!