Good morning everyone... 😆 It is with great pleasure that I have been helping this project, but lately I see that it is a little abandoned. Many people stopped participating, and the work has grown a lot. Please help, because from what I have read, our help has been very important for the discovery of new proteins that help a lot in the development of therapies to cure diseases. I appeal to users to try a little harder. Thank you all.

Well, i used to have more cpu's running Rosetta.
Now it's just one.
The lack of research updates, and not knowing what we are working on are to blame i think, well for me it is.
I also run WCG and F@h with multiple cpu's and gpu's, but there we know what we are working on and we get steady research updates.
Here everything is a black box. You get WU's, no idea what they do or for what project, you process them, and you never hear anything, that's it.
No idea what the WU's represent and a complete lack of any type of research update.
That doesn't motivate people to donate their computing resources at all.
Hence people are leaving for other communities where all that info is indeed present.
Well, that's my symplistic take at this.

tvdsluis,

You are correct that communication leaves a lot to be desired.

Sometimes they use the Twitter page, as they did recently - https://twitter.com/rosettaathome

---

Well, i used to have more cpu's running Rosetta.
Now it's just one.
The lack of research updates, and not knowing what we are working on are to blame i think, well for me it is.
I also run WCG and F@h with multiple cpu's and gpu's, but there we know what we are working on and we get steady research updates.
Here everything is a black box. You get WU's, no idea what they do or for what project, you process them, and you never hear anything, that's it.
No idea what the WU's represent and a complete lack of any type of research update.
That doesn't motivate people to donate their computing resources at all.
Hence people are leaving for other communities where all that info is indeed present.
Well, that's my symplistic take at this.

Here everything is a black box. You get WU's, no idea what they do or for what project, you process them, and you never hear anything, that's it.
No idea what the WU's represent and a complete lack of any type of research update.

This is probably by design to keep the competition from knowing exactly what we are doing and beating Rosetta to the punch. In the Science World it's not that you 'contributed' it's that you were first and did the majority of the work when it comes to getting the next grant.

ie if I told you I was was looking for grants to setup a base on the Moon you would laugh at me and say 'Nasa is already doing that' and the the grant doorway would slam shut. BUT if I didn't tell you what I was doing until I was done and everything was priced out and I told you Nasa bought my idea you would say 'WELL DONE I hope you got a boatload of money'.

This is probably by design to keep the competition from knowing exactly what we are doing and beating Rosetta to the punch. In the Science World it's not that you 'contributed' it's that you were first and did the majority of the work when it comes to getting the next grant.

It's a choice. A lot of scientific project (not only boinc project) release their source code...

---

Well, I used to have more cpu's running Rosetta.
Now it's just one.
The lack of research updates, and not knowing what we are working on are to blame I think, well for me it is.
I also run WCG and F@h with multiple cpu's and gpu's, but there we know what we are working on and we get steady research updates.
Here everything is a black box. You get WU's, no idea what they do or for what project, you process them, and you never hear anything, that's it.
No idea what the WU's represent and a complete lack of any type of research update.
That doesn't motivate people to donate their computing resources at all.
Hence people are leaving for other communities where all that info is indeed present.
Well, that's my simplistic take at this.

Blimey.
I also run WCG. I do see their updates. I'm not aware of anything they're working on or it's success or failure tbh.
It rather seems to me they started late and are taking several times longer to work on problems that've already been solved via Rosetta

It's true, Rosetta doesn't have the marketing budget on PR that IBM does, but I'm aware of their contribution to Moderna, Pfizer, Icosavax and the SK Biosciences one which seems like it'll make current vaccines look like an aspirin.
See https://twitter.com/KingLabIPD and https://twitter.com/UWproteindesign

I'll continue to run Open Pandemics on WCG, as I do all their sub-projects, but I'm not sure if I'm wasting my time for the sake of having smoke blown up my arse a lot more often.

---

Well, I used to have more cpu's running Rosetta.
Now it's just one.
The lack of research updates, and not knowing what we are working on are to blame I think, well for me it is.
I also run WCG and F@h with multiple cpu's and gpu's, but there we know what we are working on and we get steady research updates.
Here everything is a black box. You get WU's, no idea what they do or for what project, you process them, and you never hear anything, that's it.
No idea what the WU's represent and a complete lack of any type of research update.
That doesn't motivate people to donate their computing resources at all.
Hence people are leaving for other communities where all that info is indeed present.
Well, that's my simplistic take at this.

Here everything is a black box. You get WU's, no idea what they do or for what project, you process them, and you never hear anything, that's it.
No idea what the WU's represent and a complete lack of any type of research update.

This is probably by design to keep the competition from knowing exactly what we are doing and beating Rosetta to the punch. In the Science World it's not that you 'contributed' it's that you were first and did the majority of the work when it comes to getting the next grant.

ie if I told you I was was looking for grants to setup a base on the Moon you would laugh at me and say 'Nasa is already doing that' and the the grant doorway would slam shut. BUT if I didn't tell you what I was doing until I was done and everything was priced out and I told you Nasa bought my idea you would say 'WELL DONE I hope you got a boatload of money'.

Astonishing...

3 days ago, the source code for RoseTTAFold - Bakerlab's new method for protein structure prediction based on Alphafold 2 - was published for everyone else to use, with installation and usage instructions

RoseTTAFold

If I come across as annoyed, it's because I am.
There are a lot of people on here and I don't always remember everyone accurately, but my impression is that you're the guy who misses no opportunity to speculate that Rosetta seems to be running down and may close or some other complete and total nonsense like that.
And above you come out with not only some imagined gibberish about hiding what they're doing in order to harvest grant money, but what turns out to be the exact opposite of what they did a matter of hours before you posted.

If I was to characterise my dozen years following this project, it's been regular small changes for a lot of the time, but in the last year or so the project has become energised and almost supercharged.
As soon as SARS-CoV2 came about, they updated all their software developments over the previous few years following the original Alphafold 2018, focussing on the virus almost exclusively, and now published a rapid assessment of what Alphafold 2 produced, from scant information aiui.

If you have any other insights into what's going on, I'd encourage you to keep posting them, because it's quite apparent to me that I can rely on the complete opposite to be closer to reality. It's no wonder the project guys spend no time here.

---

3 days ago, the source code for RoseTTAFold - Bakerlab's new method for protein structure prediction based on Alphafold 2 - was published for everyone else to use, with installation and usage instructions

RoseTTAFold

That's a GREAT news!!!

---

That's a GREAT news!!!

Do we get to run them? The fact that they have been published for the world before we see them is curious.

---

That's a GREAT news!!!

Do we get to run them? The fact that they have been published for the world before we see them is curious.

I think it means we already did and the process is now publicly available for any lab to take the next step and run with it to make alot of money with a vaccine etc

I think it means we already did and the process is now publicly available for any lab to take the next step and run with it to make alot of money with a vaccine etc

That is OK with me, but I wonder what we will be left with? They don't tell us anything at all.

---

Over at the Robetta queue, there are some projects that were submitted for use of the RoseTTAFold method.

Kinda looks like they are comparing the results with the cameo runs (IIRC, they're like CASP but on a regular basis).

https://robetta.bakerlab.org/queue.php?id=&target=&username=&seq=&page=10

---

That's a GREAT news!!!

Do we get to run them? The fact that they have been published for the world before we see them is curious.

I think it means we already did and the process is now publicly available for any lab to take the next step and run with it to make a lot of money with a vaccine etc

Whoever turns new discoveries into a real world product that makes a significant difference will earn some money.

You make it sound like people better not discover anything groundbreakingly new that may make a big difference to the world in case someone eventually makes some money at the end.

The point isn't the money, just in case you're confused.

Honestly, the things you come out with... I despair

---

I don't know anything about what or how this project works. I just think it sounds like a good project to run.
I'm not that great on computers either. It's abundantly clear to me a lot of you guys know loads more than me.
But I'm a trier. And I read stuff. Even things that are way above my ability to understand.
I consider it a minor victory when I can read some of the more technical stuff and make it sound like coherent sentences in my head.
And I can click links and eventually work out where I'm being taken. And I can remember things - whether I understood them or not. So...

On that RoseTTAFold page I saw a link to the Robetta page and on that I saw the following text

September 15, 2020 - A deep learning based modeling method, TrRosetta, is now available as an option for modeling. For more information on this method, you can refer to the TrRosetta publication.

I've heard the name TrRosetta before, but I don't actually know what it is or what the significance is of it.

So I clicked on the TrRosetta publication link and discovered that it was published January 21, 2020 and TrRosetta stands for "transform-restrained Rosetta". I hadn't even known that before.

More particularly, I discovered it was a paper on the CASP13 (2018) results where Alphafold (Google Deepmind AI) first made an appearance and blew everyone away.
This is where ploughing through the text and hanging in there even if it's hard to take in pays off.
Bottom line is TrRosetta takes onboard everything other entrants to CASP13 reported about their methods, including Alphafold 2018, and combines it with what they already knew, resulting in an improvement on all of them.

Now, I'm not too sure about this, but when the pandemic first started there was a point in early April 2020 where everything they'd been working on here was recompiled and a new version of Rosetta came out for us to run.
We had all that trouble with permissions and outages and all sorts for a while - we even had David Kim and Brian Coventry make an appearance in these forums iirc - and among all the problems and solutions, I recall them saying the results coming back in were next level to what they were used to.
Being 3 months after that TrRosetta paper was published, I'm guessing (only guessing) that TrRosetta is what we were using here from April 2020. And 5 months later it got released to everyone else.

And further on that Robetta page it says the following:

June 15, 2021 - An improved deep learning based modeling method, RoseTTAFold, is now available. Since being benchmarked in CAMEO for the last month, RoseTTAFold has been the top ranked method

This links to a paper dated June 15 2021 that describes what it does - Accurate prediction of protein structures and interactions using a 3-track network

Alphafold 2 performance at CASP14 was reported as being so accurate it wasn't possible to be sure that Alphafold 2's results weren't actually better than the unpublished target they were trying to predict - such was the level it reached. But they haven't yet published their detailed methods beyond a generalised description of how they approached it - an approach that was completely different to the CASP13 one

It seems RoseTTAfold is a best guess at what Alphafold 2 tries to do from the limited info they have - and the limited hardware they have to run it on.
While they haven't reached Alphafold's performance by a large margin, at the same time it's already a lot better than TrRosetta. The implication being it'll improve further once more details are published about Alphafold 2.

And this time, publication of the source of RoseTTAFold for everyone else comes within a month.

There's also mention of PyRosetta on those pages.
This is another one I have no idea about until there's some link I can be pointed to that'll explain it to me in a language a 5yo can understand - because that's about my level.
Until then, I'll just consider it "a good thing".

Anyway, I just wanted to write down what I think I've learned, in case anyone else was wondering what's going on generally.
People can complain all they like about not being told anything, but it looks like it's all being explained in those places I've pointed to, among others.

If I've got anything wrong (highly likely) feel free to correct me and explain it better than me because, as I said at the top, I don't actually know anything. I just read things and understand them as best I can (or can't).

---

These tweets from the Rosetta@home twitter account don't seem to make the case that we're running TrRosetta since April/May 2020.
From what I can tell, PyRosetta is TrRosetta which we are only just beginning to run.

https://twitter.com/RosettaAtHome/status/1403057126541447170

IIRC, the Rosetta software update that occurred in April/May 2020 was really incorporating updates to the Rosetta software that hadn't yet been deployed at R@H but not TrRosetta. Something about improving the design capabilities of Rosetta for the pandemic, I recall, as Rosetta was mostly focused/built onto structure prediction.

---

I think it means we already did and the process is now publicly available for any lab to take the next step and run with it to make a lot of money with a vaccine etc

That is OK with me, but I wonder what we will be left with? They don't tell us anything at all.

The last year may've caused people to forget, but Rosetta is about protein structure prediction.
It's other projects, like WCG, to take that prediction capability and point it a specific problem.

Rosetta is about improving the engine. Where people want to drive the car it goes into is a different question.

---

The last year may've caused people to forget, but Rosetta is about protein structure prediction.
It's other projects, like WCG, to take that prediction capability and point it a specific problem.

Rosetta is about improving the engine. Where people want to drive the car it goes into is a different question.

They are doing more than that.

“We use this software called Rosetta, that’s been developed here at the U.S. and worldwide, to make up new protein nano-particles and we dial in their structures at the atomic levels. That atom here, that atom there to predictively design new nano-particles that are going to make the best possible vaccine,” said King.

https://www.king5.com/article/news/health/coronavirus/vaccine/universal-coronavirus-covid-19-vaccine-uw-medicine-research/281-58e93cf5-cd24-4355-a46d-728ded33bb75

---

These tweets from the Rosetta@home twitter account don't seem to make the case that we're running TrRosetta since April/May 2020.
From what I can tell, PyRosetta is TrRosetta which we are only just beginning to run.

https://twitter.com/RosettaAtHome/status/1403057126541447170

IIRC, the Rosetta software update that occurred in April/May 2020 was really incorporating updates to the Rosetta software that hadn't yet been deployed at R@H but not TrRosetta. Something about improving the design capabilities of Rosetta for the pandemic, I recall, as Rosetta was mostly focused/built onto structure prediction.

Straight after writing that blurb I looked at the "Rosetta and TrRosetta" thread to remind myself what was happening there.
What you've said sounds more right than what I put as they seem to be progressively adding bits of TrRosetta and I don't think it's been released to us yet.
The point stands that there's far more to come here.

Plus, I reminded myself of the image they added to one of their updates, which is worth reposting here

https://user-images.githubusercontent.com/4205776/101086651-2e5e9000-357f-11eb-880f-0f780f8f70b3.png

---

The last year may've caused people to forget, but Rosetta is about protein structure prediction.
It's other projects, like WCG, to take that prediction capability and point it a specific problem.

Rosetta is about improving the engine. Where people want to drive the car it goes into is a different question.

They are doing more than that.

“We use this software called Rosetta, that’s been developed here at the U.S. and worldwide, to make up new protein nano-particles and we dial in their structures at the atomic levels. That atom here, that atom there to predictively design new nano-particles that are going to make the best possible vaccine,” said King.

https://www.king5.com/article/news/health/coronavirus/vaccine/universal-coronavirus-covid-19-vaccine-uw-medicine-research/281-58e93cf5-cd24-4355-a46d-728ded33bb75

I don't have access to that link, but based on what you quote, yes this is the big change to the project's focus that happened in March 2020, but this is an anomaly to what Rosetta has done in its entire existence up to then due to the urgency of the pandemic. And KingLabs is a spin-off with a focus on real-world applications - one of several who submit their tasks to our queue.

If other organisations cover the real-world applications, using the Rosetta engine, that's a good thing. And Rosetta will no doubt happily return to its original purpose of improving the engine.

---

The link appears to be geographically restricted. Using a VPN connected to a US server worked.

---