• Upload from the PDBe database via PDB ID with Assembly ID (optional) or as a biological unit (recommended).
• Upload PDB or mmCIF formats.

The interface is structured with the following components. On the left, you can find a panel where switching between the Compute and Channels panels is possible. In the compute panel, you can switch the submission settings to Pore mode and Channels mode.

Pore mode - transmembrane pores

• Membrane position from the Orientations of Proteins in Membranes (OPM) database or calculated with the MEMEMBED program
• Membrane region parameter - calculate pore in the transmembrane region only
• Beta structure parameter - recommended for transmembrane regions formed by β-barrel structure

Channels mode - identification of channels and other types of pores

• Start and end point user definition.
  • list of residues (selectable from sequence or 3D structure)
  • XYZ coordinates
  • using PatternQuery language
  • on the surface using Ctrl + left mouse click (Figure 1)

• Automatic Start points definition in the deepest points within cavities -

  • Origins
  • CSA -residues from the Catalytic Site Atlas (CSA)
  • Cofactor

• Cavity parameters

  • Probe Radius (default 5 Å) - define coarseness of the surface (larger values are recommended for the detection of larger channels, e.g. 20 Å for detection of peptide channel in ribosome)
  • Interior Threshold (default 1.1 Å) - define the minimal size of tetrahedrons (smaller allows detection of channels with smaller bottlenecks)

• Channel parameters

  • Origin Radius (default 5 Å) - optimisation of starting point within the sphere
  • Surface Cover Radius (default 10 Å) - only channels with endpoints further than this parameter apart are shown
  • Weight Function selection of priority among channels
    • Voronoi scale (default) - adds to the original Length and Radius weight function and the differences between the geometrical centres of tetrahedrons and resulting channel spheres. As such, the Voronoi scale weight function prefers inward-facing tetrahedrons. Weight is calculated as W = (L/(DD+ε))(S(a)+S(b)), where L is the length of the edge, D is the distance to the closest vdW sphere, ε is a small number to avoid division by zero and S(a) and S(b) is the distance between the centre of geometrical centres of two closest tetrahedrons and resulting channel spheres.
    • Length+Radius (original in previous MOLEonline version) - the weight of edges for channels are calculated as W = L/(D*D+ε),
    • Length - weight is calculated just as W = L. This function does not take into account bottlenecks.
  • Merge Pores (disabled on default) - merge tunnels starting from the same starting point to form pores
  • Automatic Pores (disabled on default) - compute pores in all cavities

• Channel Advanced parameters

  • Bottleneck Radius (default 1.2 Å) - minimum radius of the channel
  • Bottleneck Tolerance (default 3 Å) - maximum length of channel segment narrower than bottleneck radius. It can be used for closed-channel detection
  • Max Tunnel Similarity (default 0.7) - limit of similarity of tetrahedrons between channels. A longer, similar channel is discarded.

Calculated channels are sorted according to their type in the Channel panel:

• Tunnels (from starting point to surface endpoint)
• Pores (from surface to surface through structure)
• Paths (between two defined points)

Channels are further sorted according to their cavities by their lengths.

Interactive channel visualisation runs in Mol*.

The channel profile can be seen in the bottom right corner of the webpage. Two profiles can be visualised simultaneously, showing physicochemical properties along channel length (on the x-axis):

• by colour filling (for more, see methods)

  • charge
  • hydrophobicity
  • hydropathy
  • polarity
  • lipophilicity - logP-scale and logD-scale
  • solubility - logS-scale
  • mutability

• by y-axis value

  • Radius - radius of maximum sphere in this layer.
  • Free Radius - channel radius to only main chains of the lining amino acids. It may represent the maximal radius given by the position of the protein backbone, which can be achieved by reconformation of the amino acid side chain aside from the channel or virtual glycine mutation.
  • BRadius – Radius + RMSF calculated from B-factors of residues within individual layers

Overall results for all considered channels are also summarised in Channels properties. Also, channels can be coloured according to their physicochemical properties.

Each submitted job is assigned a unique random job ID to ensure private and secure access to the results. Users can easily browse the history of their submissions by accessing the history located in the bottom left corner.

Users can compare their computed channels with the annotated channels stored in the ChannelsDB 2.0 database if such annotations are available. This feature provides valuable insights by allowing users to evaluate their findings against previously characterized channels.

The results of calculations can be downloaded in mmCIF format, which provides a detailed overview of each channel, containing various information essential for tunnel identification and analysis. It includes the ID by which the tunnel can be identified and the type of tunnel (e.g. pore, path, etc.). The calculation method and software used in the analysis, such as MOLE or Caver, are included.

Each tunnel has its profile, containing the XYZ coordinates defining the tunnel's centre and radius. Another part of the file includes the channel profile layers, which detail the individual layers of the tunnel along with their calculated physicochemical properties.

Results can be downloaded in other formats:

• Channel coordinates in mmCIF and PDB format
• PyMol, VMD, and Chimera visualisation scripts
• All information as rich JSON or ZIP (containing CSV, XML) files for further textual analysis
• Channel profiles in PNG or SVG format
• Report for individual channels can be downloaded in PDF format