README

About mCSM-metal

mCSM-metal is a tool designed to evaluate how mutations influence residue-specific metal-binding probabilities, capturing localized impacts on binding sites and broader global effects on metal coordination.

Its comprehensive workflow integrates structural mutation modeling with ESMBind, a state-of-the-art pretrained transformer model, to estimate differences in metal-binding probabilities between wild-type and mutant residues.

mCSM-metal Homepage

The mCSM-Homepage offers a user-friendly interface for researchers to navigate and submit jobs for predicting mutation-driven effects on metal-binding sites.

Welcome Section: Provides an introduction to mCSM-metal, including its purpose and applications (1)
Navigation Links: Quick access to Readme (2) and BioSig Lab (3) Homepage
Submit Prediction: The Get Started button (4) directs users to a step-by-step guide for running predictions
Free Accessibility Notice: mCSM-metal is freely accessible to all users, providing open access to its prediction framework without any usage fees (5)

Running Predictions

Clicking on the Get Started button on the homepage directs users to the prediction page where they can Provide Input Data and run mutation based predictions.

The Provide Input Data page offers two options for users to query the server with their protein of interest, allowing them to specify structural details and mutations for analysis.

Input Requirements

Option 1: Provide Input Text panel (A)

PDB code and Chain ID: Submit a valid PDB four-letter code with the chain identifier (e.g., 1qt1_A) (1)
Mutation(s): Specify mutation details manually as a text string in the format: wild-type residue (one-letter code) + position + mutant residue (e.g., D244A) (2)
A list of multiple mutations can be submiited to be processed in batch following the mutation format as mentioned above and comma-separated (e.g., D244A,D286G)
Click Submit (3)

Option 2: Upload File panel (B)

Upload a single CIF file: Upload a structure file in the Crystallographic Information File (CIF) format by clicking on the Choose File button (4). Ensure that there are no special characters in the file name.
Chain ID: Specify the chain ID in the text box where mutation should be introduced (5)
Specify mutation details manually as a text string in the format: wild-type residue (one-letter code) + position + mutant residue (e.g., D244A) (6)
A list of multiple mutations can be submitted to be processed in batch following the mutation format as mentioned above and comma-separated (e.g., D244A,D286G)
Click Submit (7)

Pre-calculated Example (C) The page also provides an example based on the Xylose Isomerase Protein (PDB: 1qt1_A), which mutates aspartic acid (D) to alanine (A) at position 244. TThe corresponding results can be accessed through the provided link (8).

Note: This resource requires complete protein structures. A protein structure with missing regions at the start or end will still pass the validations carried out in this resource, however the predictions might be different compared to a complete structure. A structure with region missing in the middle will not pass the validation stage and will not be processed by this resource.

Visualizing and Interpreting Results

Upon landing on the result page, users will initially see four empty panels. These panels will populate with detailed results once a residue is selected using the Feature Viewer (D). The result page, generated from either Option 1 or Option 2 inputs (mentioned above), provides a similar comprehensive view of mutation-induced changes in metal-binding probabilities through these following key components:

Protein Structure Viewer (A)
- The prediction title displays the PDB ID submitted by the user and the selected ion being predicted (e.g., Prediction on 1qt1_A for ZN ion) (1)
- The structure viewer is integrated with Mol, this 3D viewer (2) enables visualization of the wild-type structure, annotated to reflect per-residue differences in metal-binding probabilities with a red-white-blue color gradient. The gradient bar is shown at the top right of the viewer (3)*
Color scale interpretation:
```
  Red: Increased metal-binding probability of MT (positive change)
```
White: No significant change Blue: Decreased metal-binding probability (negative change)
- Users can select one of seven ions from the drop-down menu (4), with the following abbreviations: CA (Calcium), CO (Cobalt), CU (Copper), FE (Iron), MG (Magnesium), MN (Manganese), and ZN (Zinc). Once an ion is selected, the structure automatically updates to display the predictions for that ion and the prediction title is updated
Residue Summary (B)

Displays details for the selected residue, including: * Residue (position): e.g., 244 * Mutation: e.g., D244A * Ion coordination status: True/False * Calculated probability difference (ΔP_norm, normalized between 0 and 1) * Category assignment based on ΔP_norm values (refer to Table 1)

Table 1. Categorization of binding probability change based on scaled ΔP_norm

Scaled ΔP_norm Value Category Label

ΔP_norm ≥ 0.75 Strong increase

0.55 ≤ ΔP_norm < 0.75 Increase

0.5 ≤ ΔP_norm < 0.55 Slight increase

0.45 ≤ ΔP_norm < 0.5 Slight decrease

0.25 ≤ ΔP_norm < 0.45 Decrease

ΔP_norm < 0.25 Strong decrease

Scaled ΔP_norm Value	Category Label
ΔP_norm ≥ 0.75	Strong increase
0.55 ≤ ΔP_norm < 0.75	Increase
0.5 ≤ ΔP_norm < 0.55	Slight increase
0.45 ≤ ΔP_norm < 0.5	Slight decrease
0.25 ≤ ΔP_norm < 0.45	Decrease
ΔP_norm < 0.25	Strong decrease

Distance Summary Plots (C)

Displays two plots to interpret the proximity of the selected residue

Line Plot: Displays distances between the selected residue and other residues across the structure (5) Bar Plot: Shows the frequency distribution of distances, indicating whether the residue is buried or exposed residues (6)
Feature Viewer (D)

This is an interactive sequence and structure visualization using RCSB-Saguaro, featuring:
- Track 1: Wildtype sequence in one letter amino acid code
- Track 2: Mutated sequence in one letter amino acid code
- Track 3: Position of Mutation(s), marked with red dots
- Track 4: Selected metal-Coordinating Residues, highlighted with blue dots
- Track 5: Scaled Probability changes per residue, shown in a light green bar chart
Most residues show minimal or no change in binding probability upon mutation, leveling around the 0.5 mark, while those with significant alterations deviate from this value.

Exporting Results

Results can be exported as CSV files by clicking the Download Probabilities (csv) button at the bottom-left of the results page (7)

API Access

mCSM-metal offers an API to submit jobs and retrieve results programmatically, providing flexibility for automated or large-scale analysis.

Job Submission

Arguments:
- PDB accession code and chain ID: Provide input using a four-character PDB code with chain identifier (e.g. 1qt1_A)
- Mutation: Provide details of the mutant(s) in the following format: amino acid from + residue number + amino acid to. Multiple mutations can be submitted, separated by commas (e.g. D244A, D286G)
Result:
- Job ID: A unique identifier is returned for each successful submission. The job ID can be used to track and retrieve the results.
Example: Submitting a Job

curl -X POST \
     -H "Content-Type: application/json" \
     -d '{"text_input": "1qt1_A", "mutations_input": "D244A,D286G"}' \
     https://biosig.lab.uq.edu.au/mcsm_metal/api/submit

Retrieving Results

Run the following command to retrieve results for any submitted job_id:

curl -X GET https://biosig.lab.uq.edu.au/mcsm_metal/api/status/<job_id>

API Response Schema

The API returns a JSON object with the following structure:

{
  "results": [
    {
      "CA": "float",      // Change in binding probability for Calcium
      "CO": "float",      // Change in binding probability for Cobalt
      "CU": "float",      // Change in binding probability for Copper
      "FE": "float",      // Change in binding probability for Iron
      "MG": "float",      // Change in binding probability for Magnesium
      "MN": "float",      // Change in binding probability for Manganese
      "Residue_ID": "integer",  // Residue position identifier (0-indexed)
      "ZN": "float"       // Change in binding probability for Zinc
    }
  ],
  "status": "string"  // Status of the request (e.g., "completed")
}

Field Descriptions:

results: A list of objects, each containing:
- Residue_ID: The residue position in the protein.
- CA, CO, CU, FE, MG, MN, ZN: The computed change in metal-binding probability for each ion (floating-point values).
status: A string indicating the completion status of the API request (e.g., "completed").

Example Applications

Understanding the effect of mutations on metalloproteins.
Studying mutation-induced loss of enzymatic activity.
Designing metal-stabilized protein structures in drug discovery

Contact and Support

For assistance or to report issues, please contact us via email or our Group Website. When reporting a job issue, include input details and the job identifier for quicker resolution.