Predicting ion mobility collision cross sections using projection approximation with ROSIE-PARCS webserver

Abstract

Ion mobility coupled to mass spectrometry informs on the shape and size of protein structures in the form of a collision cross section (CCSIM). Although there are several computational methods for predicting CCSIM based on protein structures, including our previously developed projection approximation using rough circular shapes (PARCS), the process usually requires prior experience with the command-line interface. To overcome this challenge, here we present a web application on the Rosetta Online Server that Includes Everyone (ROSIE) webserver to predict CCSIM from protein structure using projection approximation with PARCS. In this web interface, the user is only required to provide one or more PDB files as input. Results from our case studies suggest that CCSIM predictions (with ROSIE-PARCS) are highly accurate with an average error of 6.12%. Furthermore, the absolute difference between CCSIM and CCSPARCS can help in distinguishing accurate from inaccurate AlphaFold2 protein structure predictions. ROSIE-PARCS is designed with a user-friendly interface, is available publicly and is free to use. The ROSIE-PARCS web interface is supported by all major web browsers and can be accessed via this link (https://rosie.graylab.jhu.edu).

Keywords: AlphaFold2; collision cross section; ion mobility; mass spectrometry; protein structure prediction; webserver.

Figure 1

Illustration of the PARCS algorithm in Rosetta. In PARCS, each protein atom is estimated as a rough circle using a 9-point approximation based on the respective atomic and buffer gas radii. The PARCS projection representation is obtained after iterating through every single atom in the structure.

Figure 2

Input page of the ROSIE-PARCS webpage. PDB file(s) can be dragged into the drop box. Alternatively, users can use the ‘Browse’ button (shown within the back dashed box) to navigate their filesystem and/or directly obtain a structure from the PDB by providing the PDB code and pressing the download button (as indicated by the green dashed box). The number of residues, atoms and the chain identifiers are displayed for successfully uploaded PDB files (orange dashed box). The two input parameters, ‘number of random rotations’ and ‘probe radius in Angstroms’ are shown in the red and blue dashed boxes.

Figure 3

Results page of the ROSIE-PARCS application. Successful completion of the CCS_PARCS calculation is indicated by the green ‘finished’ State. CCS_PARCS values are displayed in a tabular format. The table headers are labeled as ‘File Name’ and ‘CCS Value (Å^2)’. The red and blue box are highlighting the ‘Download Results’ and copy to clipboard buttons, respectively.

Figure 4

ROSIE-PARCS (blue) and CLI-PARCS (orange) produce virtually identical CCS_PARCS predictions at various parameter settings. (A) CCS_PARCS (normalized by sequence length) for both ROSIE-PARCS and CLI-PARCS at probe radii between 1.0 and 2.0 Å and at a fixed NRRs of 300. (B) CCS_PARCS (normalized by sequence length) for both ROSIE-PARCS and CLI-PARCS using 100–600 random rotations (at a fixed probe radius of 1.0 Å).

Figure 5

Comparison of CCS_PARCS from ROSIE-PARCS to CCS_IM. (A) The CCS values without normalization are shown. (B) CCS values normalized by the sequence length for all proteins in the CCS_IM dataset are shown.

Figure 6

Analysis of average pLDDT, ΔCCS and RMSD for all predicted structures from AF2. Comparison of (A) the AF2 confidence metric (average pLDDT) and (B) the absolute difference between CCS_IM and CCS_PARCS (ΔCCS) against the RMSD of the predicted structures.