What is Collisional Cross-Sectional (CCS) ?

Mass Spectrometry (MS) allows separation of proteins based on their mass-to-charge ratio, but does not provide an indication as to the shape of the protein.

Ion Mobility Mass Spectrometry (IM-MS) first lets the protein move through a buffer gas. Collisions with this buffer gas slow down the protein in a size-dependent, rather than mass-dependent, fashion, allowing separation of conformations and improved resolution.

CCS is the measure of how much the protein is slowed down and is roughly proportional to the mean cross-sectional area of the protein, expanded by the diameter of the buffer gas.

What Is ProteoTwirl-CCS?

ProteoTwirl-CCS is a free web service, which allows a prediction of CCS from a protein structure. ProteoTwirl-CCS will calculate CCS using one of ProteoTwirl's methods, either O-DISC or CA-TRAACE-corrected O-DISC (see below), and provide a range of visualizations to analyze the data.

What is O-DISC?

Overlapping Disk Interior from Subdivided Cells (O-DISC) is a projection approximation (PA) method for CCS prediction. This looks at the protein from a range of different angles and calculates the area of the silhouette of the protein as seen from that angle. The CCS value is then the average of this across all angles.

O-DISC is quite fast, often requiring tens of milliseconds per protein. However, like many PA methods, O-DISC has a consistent underestimation of the experimental CCS value. CA-TRAACE is thus a method designed to account for this.

What is CA-TRAACE?

Curvature Adaptive Topology Refinement for Area Adjustment from Collision Effects (CA-TRAACE) is a method designed to adjust the CCS value from projection approximation (PA) methods like O-DISC to counteract the underestimation of the method.

For each angle observed by O-DISC, CA-TRAACE will calculate a multiplier, which will get multiplied by the O-DISC area to get the CCS contribution from that orientation. The final CCS value is then the mean of these contributions.

CA-TRAACE essentially works by simulating rays colliding with specific atoms on the protein surface and calculating how often such a ray will get deflected in a direction where it would collide with an additional atom. This corrects for bumps on the protein surface. CA-TRAACE, however, does not account for long-range interactions that also can affect CCS.

Timeout and Usage Limits

ProteoTwirl-CCS allows a job to take a maximum of a few minutes to calculate, and the job will be terminated afterwards. If your protein keeps timing out, we recommend disabling the visualizations, as these take the majority of the compute time.

There is also a limit on upload size. If your protein exceeds this, try removing hydrogens from it if they are present. If it is still too large, the protein cannot be predicted using the free online web service. However, you can download and run the ProteoTwirl-CCS Docker container locally and disable the upload limit, or alternatively you can use the ProteoTwirl Python package to analyze the data.

We do ask users not to submit a large number of jobs, especially if there is a queue, and, ideally, to submit only one job at a time and not to submit another before that one has finished. If you desire to calculate CCS for large batches of structures or frames of molecular dynamics simulations, we recommend using the ProteoTwirl Python package instead.

ProteoTwirl vs ProteoTwirl-CCS

ProteoTwirl-CCS is a free online web service for CCS calculation, using ProteoTwirl (the Python package) for the underlying calculations. ProteoTwirl allows calculation of a range of orientation-dependent properties, not only CCS values, allowing analysis of the hydrophobicity of the protein surface or the performance of binding pocket analysis.

Source Code & Links

ProteoTwirl-CCS and ProteoTwirl are both free and open-source

Citing ProteoTwirl-CCS

If ProteoTwirl-CCS was used in a place where appropriate, please cite it as the ProteoTwirl paper [COMING SOON].

Acknowledgement

Location estimates use a locally installed database. When DB-IP data is installed: IP Geolocation by DB-IP. With ProteoTwirl-CCS implemented using FastAPI as the backend.

ProteoTwirl uses MDAnalysis and PDB-Fixer to read and parse input files, and uses CartoPy for cartographic visualizations and ThreeJS for interactive 3D visualization. Note that both cartographic and 3D visualization standard visualization function in ProteoTwirl.

Contact

Contact: proteotwirl-ccs@mol.tools.