| 
							 | 
						--- | 
					
					
						
						| 
							 | 
						dataset_info: | 
					
					
						
						| 
							 | 
						  features: | 
					
					
						
						| 
							 | 
						  - name: input_ids | 
					
					
						
						| 
							 | 
						    sequence: int32 | 
					
					
						
						| 
							 | 
						  - name: coords | 
					
					
						
						| 
							 | 
						    sequence: | 
					
					
						
						| 
							 | 
						      sequence: float64 | 
					
					
						
						| 
							 | 
						  - name: labels | 
					
					
						
						| 
							 | 
						    dtype: float64 | 
					
					
						
						| 
							 | 
						  splits: | 
					
					
						
						| 
							 | 
						  - name: train | 
					
					
						
						| 
							 | 
						    num_bytes: 1434933392 | 
					
					
						
						| 
							 | 
						    num_examples: 12479 | 
					
					
						
						| 
							 | 
						  - name: val | 
					
					
						
						| 
							 | 
						    num_bytes: 251648000 | 
					
					
						
						| 
							 | 
						    num_examples: 4000 | 
					
					
						
						| 
							 | 
						  - name: test | 
					
					
						
						| 
							 | 
						    num_bytes: 249528000 | 
					
					
						
						| 
							 | 
						    num_examples: 4000 | 
					
					
						
						| 
							 | 
						  download_size: 718847118 | 
					
					
						
						| 
							 | 
						  dataset_size: 1936109392 | 
					
					
						
						| 
							 | 
						configs: | 
					
					
						
						| 
							 | 
						- config_name: default | 
					
					
						
						| 
							 | 
						  data_files: | 
					
					
						
						| 
							 | 
						  - split: train | 
					
					
						
						| 
							 | 
						    path: data/train-* | 
					
					
						
						| 
							 | 
						  - split: val | 
					
					
						
						| 
							 | 
						    path: data/val-* | 
					
					
						
						| 
							 | 
						  - split: test | 
					
					
						
						| 
							 | 
						    path: data/test-* | 
					
					
						
						| 
							 | 
						--- | 
					
					
						
						| 
							 | 
						 | 
					
					
						
						| 
							 | 
						# RSR: RNA Structure Ranking | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						## Overview | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						This task relates to predicting the three-dimensional structure of an RNA | 
					
					
						
						| 
							 | 
						molecule, given its sequence.  A total of 21 RNAs are included, which consist of | 
					
					
						
						| 
							 | 
						the first 21 RNAs from the RNA-Puzzles competition (Cruz et al., 2011). | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						We phrase this problem as candidate ranking.  For each RNA, we generate | 
					
					
						
						| 
							 | 
						candidate structural models using FARFAR2 (Watkins et al., 2020) and calculate | 
					
					
						
						| 
							 | 
						each candidate's atoms' root mean squared deviation (RMSD) to the experimentally | 
					
					
						
						| 
							 | 
						determined structure. | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						## Datasets | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						- splits: | 
					
					
						
						| 
							 | 
						   - candidates-split-by-time: Oldest RNAs are in training set, next oldest are in validation, and newest are in testing set. | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						## Additional Information | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						The entry ['scores'] contains a dictionary with the following important entries: | 
					
					
						
						| 
							 | 
						- 'rms': RMSD from the experimentally determined structure. | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						Many of the other scores are documented in [https://daslab.stanford.edu/site_data/pub_pdf/2017_Alford_JCTC.pdf](https://daslab.stanford.edu/site_data/pub_pdf/2017_Alford_JCTC.pdf) | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						The ensemble denotes the | 
					
					
						
						| 
							 | 
						number of the RNA Puzzle, and the subunit denotes the index of the candidate | 
					
					
						
						| 
							 | 
						structural model. | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						## Citation Information | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						``` | 
					
					
						
						| 
							 | 
						@article{townshend2020atom3d, | 
					
					
						
						| 
							 | 
						  title={Atom3d: Tasks on molecules in three dimensions}, | 
					
					
						
						| 
							 | 
						  author={Townshend, Raphael JL and V{\"o}gele, Martin and Suriana, Patricia and Derry, Alexander and Powers, Alexander and Laloudakis, Yianni and Balachandar, Sidhika and Jing, Bowen and Anderson, Brandon and Eismann, Stephan and others}, | 
					
					
						
						| 
							 | 
						  journal={arXiv preprint arXiv:2012.04035}, | 
					
					
						
						| 
							 | 
						  year={2020} | 
					
					
						
						| 
							 | 
						} | 
					
					
						
						| 
							 | 
						``` | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						``` | 
					
					
						
						| 
							 | 
						@article{cruz2012rna, | 
					
					
						
						| 
							 | 
						  title={RNA-Puzzles: a CASP-like evaluation of RNA three-dimensional structure prediction}, | 
					
					
						
						| 
							 | 
						  author={Cruz, Jos{\'e} Almeida and Blanchet, Marc-Fr{\'e}d{\'e}rick and Boniecki, Michal and Bujnicki, Janusz M and Chen, Shi-Jie and Cao, Song and Das, Rhiju and Ding, Feng and Dokholyan, Nikolay V and Flores, Samuel Coulbourn and others}, | 
					
					
						
						| 
							 | 
						  journal={Rna}, | 
					
					
						
						| 
							 | 
						  volume={18}, | 
					
					
						
						| 
							 | 
						  number={4}, | 
					
					
						
						| 
							 | 
						  pages={610--625}, | 
					
					
						
						| 
							 | 
						  year={2012}, | 
					
					
						
						| 
							 | 
						  publisher={Cold Spring Harbor Lab} | 
					
					
						
						| 
							 | 
						} | 
					
					
						
						| 
							 | 
						``` | 
					
					
						
						| 
							 | 
						
 | 
					
					
						
						| 
							 | 
						``` | 
					
					
						
						| 
							 | 
						@article{watkins2020farfar2, | 
					
					
						
						| 
							 | 
						  title={FARFAR2: improved de novo rosetta prediction of complex global RNA folds}, | 
					
					
						
						| 
							 | 
						  author={Watkins, Andrew Martin and Rangan, Ramya and Das, Rhiju}, | 
					
					
						
						| 
							 | 
						  journal={Structure}, | 
					
					
						
						| 
							 | 
						  volume={28}, | 
					
					
						
						| 
							 | 
						  number={8}, | 
					
					
						
						| 
							 | 
						  pages={963--976}, | 
					
					
						
						| 
							 | 
						  year={2020}, | 
					
					
						
						| 
							 | 
						  publisher={Elsevier} | 
					
					
						
						| 
							 | 
						} | 
					
					
						
						| 
							 | 
						``` | 
					
					
						
						| 
							 | 
						
 |