FRET-guided modeling of nucleic acids

Note

Please cite: Fabio D Steffen, Richard A Cunha, Roland K O Sigel, Richard Börner, FRET-guided modeling of nucleic acids, Nucleic Acids Research, 2024, https://doi.org/10.1093/nar/gkae496

This repository contains structural models of the proof-of-concept de novo modeling trial of a manganese(II) sensing riboswitch (PDB 6n2v). The modeling code can be found here.

The resulting models are organized as follows:

• 1_scored_by_rosetta_energy/ contains the top 10 models scored by the Rosetta energy function from FARFAR2 (Watkins, Structure 2020) was well as donor and acceptor accessible-contact volumes (ACV). The predicted mean FRET values computed by FRETraj and the Rosetta energy scores are listed in top10_rosetta_energy_FRET_prediction.csv

state	FRET	name	Rosetta_score
1	0.79	S_000292_003	-193.882
2	0.1	S_000014_004	-188.55
3	0.09	S_000290_008	-186.863
4	0.07	S_000059_012	-185.191
5	0.19	S_000083_009	-185.004
...

Visualize the ACVs on the structural models by running the following Python script from the PyMOL command line

run top10_rosetta_energy.py

---

• 2_scored_by_FRET_RMSD/ contains models filtered by a FRET treshold >0.4 and a RMSD < 15 A. All models are superimposed and aligned to the crystal structure (PDB 6n2v). The best model (no. 21) has a RMSD of 8.6 A to the crystal structure. The calculated RMSDs and Rosetta energy scores are listed in models_scored_by_FRET_and_RMSD.csv

state	RMSD	name	Rosetta_score
1	11.70476341	S_000355_008	-184.459
2	11.52333355	S_000086_010	-175.494
3	12.03283024	S_000054_007	-173.63
...
21	8.576731682	S_000273_005	-157.595
...

Visualize the overlay of the low RMSD models

run 2.1_lowest_RMSD_model_overlay.py

Compare the individual models to the crystal structure

run 2.2_lowest_RMSD_model_to_crystal.py