Blind Prediction of Complex Water and Ion Ensembles Around RNA in CASP16 - LCBio group (TS189)
Description
This dataset comprises the molecular dynamics (MD) simulation data generated by the LCBio group (TS189) for the CASP16 RNA-water/ion prediction challenge. The collection provides the underlying simulation data used to derive the submitted predictions for target R1260. The data were generated using a simulation pipeline performed with Amber 22 (pmemd.cuda), focusing on: - System Preparation: System topology was generated using Amber's tleap, solvating the RNA (based on PDB 7EZ0) in a truncated octahedral box with a 10 Å buffer and neutralizing it with Na+ ions. - Simulation Protocol: The process included heating from 100K to 300K over 500ps with backbone restraints, followed by a four-phase NVT equilibration and a 10ns production run. - Trajectory Analysis: 1000 frames were selected for the final analysis. While the final submitted predictions had water molecules beyond 5 Å from the RNA stripped, the trajectories provided in this repository retain all water molecules and ions. This dataset explicitly supports the findings reported in the article "Blind Prediction of Complex Water and Ion Ensembles Around RNA in CASP16" and provides the raw molecular dynamics trajectories used to generate the submitted water/ion predictions. It serves as a valuable resource for researchers interested in benchmarking solvation prediction methods, analyzing RNA-ion interactions, or studying the structural dynamics of the specific targets involved in the CASP16 challenge.
Files
Steps to reproduce
1. System Preparation: The initial RNA model (PDB 7EZ0) was minimized using QRNAS. System topology was generated using Amber's tleap via the provided `run.leap` script, which solvated the RNA in a truncated octahedral box with a 10 Å buffer and neutralized it with Na+ ions. 2. Heating: The system was gradually heated from 100 K to 300 K over 500 ps (250,000 steps) with a time step of 2 fs. Positional restraints (20.0 kcal/mol·Å²) were maintained on the RNA backbone. A Langevin thermostat (`ntt=3`) with a collision frequency `gamma_ln` of 5.0 ps⁻¹ was used. The nonbonded cutoff was 12 Å, and SHAKE constraints were applied to bonds involving hydrogen (`ntc=2`, `ntf=2`). 3. Equilibration: Following heating, the system underwent a four-phase equilibration process under NVT conditions. 4. Production & Selection: A 10 ns production run was performed using Amber 22 (pmemd.cuda). 1000 models were selected from the trajectory. For the final predictions, water molecules within 5 Å distance from at least one of the atoms from the RNA chain were retained. While the final submitted predictions had water molecules beyond 5 Å from the RNA stripped, the trajectories provided in this repository retain all water molecules and ions. For more comprehensive details on the simulation parameters and protocols, please refer to the associated research article.
Institutions
- University of WarsawMazovia, Warsaw