{ "id": 475103, "title": "Seminar on Ribonanza outcome: Tuesday, 13 February, 2024", "url": "/competitions/stanford-ribonanza-rna-folding/discussion/475103", "author_name": "Rhiju Das", "post_date": "2024-02-07T04:41:53.493000", "votes": 4, "comment_count": 0, "views": 0, "content": "

Hi everyone! A preprint on our Ribonanza work, along with curated models and data are almost done -- we should have everything out soon.

\n

In the meanwhile, if you're interested in hearing about substantial progress enabled by this Kaggle challenge, please check out host and Kaggle grandmaster @shujun717 next Tuesday, 13 February, in a seminar hosted by the CASP RNA Special Interest Group:

\n

https://hackmd.io/@mmagnus/Skr0wPljp

\n

Ribonanza: deep learning of RNA structure through dual crowdsourcing

\n

This is follow-up to our previous talk of Rhiju Das (https://www.youtube.com/watch?v=R6-MwrGkj7M&)

\n

Zoom link Tuesday Feb 13th Pacific Time 8 am / Eastern Time 11 am / Central European Time: 5 pm / China Standard Time: 11 pm

\n

Zoom link: https://stanford.zoom.us/j/93445935624?pwd=K0VUWk0zaVNMZlU1U0xUMS8vSWUwZz09

\n

Abstract
\nPrediction of RNA structure from sequence remains an unsolved problem, and progress has been slowed by a paucity of experimental data. Here, we present Ribonanza, a dataset of chemical mapping measurements on two million diverse RNA sequences collected through Eterna and other crowdsourced initiatives. The Ribonanza measurements enabled solicitation, training, and prospective evaluation of diverse deep neural networks through a Kaggle challenge, followed by distillation into a single, self-contained model called RibonanzaNet. When fine tuned on previous datasets, RibonanzaNet achieves state-of-the-art performance in modeling RNA chemical stability and RNA secondary structure. The latter enables three-dimensional structure modeling of RNA-only targets from the recent CASP trials with accuracy approaching human expert groups.

\n

Bio
\nShujun He is a PhD student in chemical engineering at Texas A&M University. Interested in applying deep learning to scientific problems. Ranked 17/194,534 (top 0.008%) globally on Kaggle (https://www.kaggle.com/shujun717), the biggest machine learning competition platform

", "messages": [ { "id": 2640764, "postDate": "2024-02-07T04:41:53.493Z", "content": "

Hi everyone! A preprint on our Ribonanza work, along with curated models and data are almost done -- we should have everything out soon.

\n

In the meanwhile, if you're interested in hearing about substantial progress enabled by this Kaggle challenge, please check out host and Kaggle grandmaster @shujun717 next Tuesday, 13 February, in a seminar hosted by the CASP RNA Special Interest Group:

\n

https://hackmd.io/@mmagnus/Skr0wPljp

\n

Ribonanza: deep learning of RNA structure through dual crowdsourcing

\n

This is follow-up to our previous talk of Rhiju Das (https://www.youtube.com/watch?v=R6-MwrGkj7M&)

\n

Zoom link Tuesday Feb 13th Pacific Time 8 am / Eastern Time 11 am / Central European Time: 5 pm / China Standard Time: 11 pm

\n

Zoom link: https://stanford.zoom.us/j/93445935624?pwd=K0VUWk0zaVNMZlU1U0xUMS8vSWUwZz09

\n

Abstract
\nPrediction of RNA structure from sequence remains an unsolved problem, and progress has been slowed by a paucity of experimental data. Here, we present Ribonanza, a dataset of chemical mapping measurements on two million diverse RNA sequences collected through Eterna and other crowdsourced initiatives. The Ribonanza measurements enabled solicitation, training, and prospective evaluation of diverse deep neural networks through a Kaggle challenge, followed by distillation into a single, self-contained model called RibonanzaNet. When fine tuned on previous datasets, RibonanzaNet achieves state-of-the-art performance in modeling RNA chemical stability and RNA secondary structure. The latter enables three-dimensional structure modeling of RNA-only targets from the recent CASP trials with accuracy approaching human expert groups.

\n

Bio
\nShujun He is a PhD student in chemical engineering at Texas A&M University. Interested in applying deep learning to scientific problems. Ranked 17/194,534 (top 0.008%) globally on Kaggle (https://www.kaggle.com/shujun717), the biggest machine learning competition platform

", "rawMarkdown": "Hi everyone! A preprint on our Ribonanza work, along with curated models and data are almost done -- we should have everything out soon.\n\nIn the meanwhile, if you're interested in hearing about substantial progress enabled by this Kaggle challenge, please check out host and Kaggle grandmaster @shujun717 next Tuesday, 13 February, in a seminar hosted by the CASP RNA Special Interest Group:\n\n\n[https://hackmd.io/@mmagnus/Skr0wPljp](https://hackmd.io/@mmagnus/Skr0wPljp)\n\n**Ribonanza: deep learning of RNA structure through dual crowdsourcing**\n\nThis is follow-up to our previous talk of Rhiju Das (https://www.youtube.com/watch?v=R6-MwrGkj7M&)\n\nZoom link Tuesday Feb 13th Pacific Time 8 am / Eastern Time 11 am / Central European Time: 5 pm / China Standard Time: 11 pm\n\n\nZoom link: https://stanford.zoom.us/j/93445935624?pwd=K0VUWk0zaVNMZlU1U0xUMS8vSWUwZz09\n\n**Abstract**\nPrediction of RNA structure from sequence remains an unsolved problem, and progress has been slowed by a paucity of experimental data. Here, we present Ribonanza, a dataset of chemical mapping measurements on two million diverse RNA sequences collected through Eterna and other crowdsourced initiatives. The Ribonanza measurements enabled solicitation, training, and prospective evaluation of diverse deep neural networks through a Kaggle challenge, followed by distillation into a single, self-contained model called RibonanzaNet. When fine tuned on previous datasets, RibonanzaNet achieves state-of-the-art performance in modeling RNA chemical stability and RNA secondary structure. The latter enables three-dimensional structure modeling of RNA-only targets from the recent CASP trials with accuracy approaching human expert groups.\n\n**Bio**\nShujun He is a PhD student in chemical engineering at Texas A&M University. Interested in applying deep learning to scientific problems. Ranked 17/194,534 (top 0.008%) globally on Kaggle (https://www.kaggle.com/shujun717), the biggest machine learning competition platform\n", "votes": 4 } ], "comments": [], "raw_markdown_by_id": { "2640764": "Hi everyone! A preprint on our Ribonanza work, along with curated models and data are almost done -- we should have everything out soon.\n\nIn the meanwhile, if you're interested in hearing about substantial progress enabled by this Kaggle challenge, please check out host and Kaggle grandmaster @shujun717 next Tuesday, 13 February, in a seminar hosted by the CASP RNA Special Interest Group:\n\n\n[https://hackmd.io/@mmagnus/Skr0wPljp](https://hackmd.io/@mmagnus/Skr0wPljp)\n\n**Ribonanza: deep learning of RNA structure through dual crowdsourcing**\n\nThis is follow-up to our previous talk of Rhiju Das (https://www.youtube.com/watch?v=R6-MwrGkj7M&)\n\nZoom link Tuesday Feb 13th Pacific Time 8 am / Eastern Time 11 am / Central European Time: 5 pm / China Standard Time: 11 pm\n\n\nZoom link: https://stanford.zoom.us/j/93445935624?pwd=K0VUWk0zaVNMZlU1U0xUMS8vSWUwZz09\n\n**Abstract**\nPrediction of RNA structure from sequence remains an unsolved problem, and progress has been slowed by a paucity of experimental data. Here, we present Ribonanza, a dataset of chemical mapping measurements on two million diverse RNA sequences collected through Eterna and other crowdsourced initiatives. The Ribonanza measurements enabled solicitation, training, and prospective evaluation of diverse deep neural networks through a Kaggle challenge, followed by distillation into a single, self-contained model called RibonanzaNet. When fine tuned on previous datasets, RibonanzaNet achieves state-of-the-art performance in modeling RNA chemical stability and RNA secondary structure. The latter enables three-dimensional structure modeling of RNA-only targets from the recent CASP trials with accuracy approaching human expert groups.\n\n**Bio**\nShujun He is a PhD student in chemical engineering at Texas A&M University. Interested in applying deep learning to scientific problems. Ranked 17/194,534 (top 0.008%) globally on Kaggle (https://www.kaggle.com/shujun717), the biggest machine learning competition platform\n" } }