I want to plot a xyz graph and also get clusters from the array of my data; can someone help please. Atoms are linked by indices. (attaching the sheet for reference)

I have combined the information from the bonds file and xyz file and want to plot a graph with various clusters (based on how the indices are connected). Is there a code to get this done with?

[Image Analyst]

fileName = 'Testfile3.xlsx'
data = readmatrix(fileName);
whos data

'Testfile3.xlsx'

Name Size Bytes Class Attributes

data 4000x23 736000 double

What do the 4000 rows represent? What do the 23 columns represent?

[ANINDYA GANGULY]

4000 rows are the 4000 atoms (2000 C and 2000 H) and the 23 columns have the position of the atoms along the x, y and z coordinates, then the heading of the 4000 atoms listed; how the indices are connected to each other (defining clusters); bond order (bo) values, bond order of the indexed atom; charge of the atom; lone pair of the atom.

If you open the .xlsx file; all the informatiom is provided in the topmost row.

Regards

J

[ANINDYA GANGULY]

And id refers to the indices of the atoms. Some indices (C) are connected 1, some with 2, some with 3 atoms (can be H/C) and so on.

Regards

J

Answer 1

[Chetan]

Hi @ANINDYA GANGULY

I understand you wish to plot a 3D graph for clusters from an array of data. To visualize a 3D graph of atoms and identify clusters based on the connectivity defined by indices, MATLAB's graph and clustering functions can be utilized.

Step 1: Import Data

filename = 'Testfile3.xlsx'; % Update this with the actual file name
data = readtable(filename);

Step 2: Create Graph

Assuming the `id` columns represent the atom indices and `id_1` to `id_5` signify connections (with zeros indicating no connections), you can construct an edge list and create a graph.

% Initialize an empty array for edges
edges = [];
% Assuming 'id_of_atom' is the unique identifier for atoms
% and connections are in 'id_1' to 'id_5' columns
for i = 1:30
    atomId = data.idOfAtom(i);
    for j = 11:15 % Assuming id_1 to id_5 are in columns 11 to 15
        connectedAtomId = data{i,j};
        if ~isnan(connectedAtomId) % Check if there is a connection
            edges = [edges; atomId, connectedAtomId];
        end
    end
end

% Create the graph
G = graph(edges(:,1),edges(:,2));

Step 3: Plotting

To plot the graph in 3D, the x, y, and z coordinates of each atom are required.

figure;
hold on;
% Plot edges
for i = 1:size(edges,1)
    atomIndex1 = find(data.idOfAtom == edges(i,1));
    atomIndex2 = find(data.idOfAtom == edges(i,2));
    
    x = [data.x_axis(atomIndex1); data.x_axis(atomIndex2)];
    y = [data.y_axis(atomIndex1); data.y_axis(atomIndex2)];
    z = [data.z_axis(atomIndex1); data.z_axis(atomIndex2)];
    plot3(x, y, z, 'k-'); % 'k-' for black lines
end
% Plot atoms on top of the edges
scatter3(data.x_axis, data.y_axis, data.z_axis, 36, 'filled');
xlabel('X-axis');
ylabel('Y-axis');
zlabel('Z-axis');
title('3D Graph of Atoms and Bonds');
hold off;

Step 4: Clustering

To identify clusters based on connectivity:

% Find connected components (clusters)
clusters = conncomp(G);

% Number of clusters
numClusters = max(clusters);

% Display number of clusters
disp(['Number of clusters: ', num2str(numClusters)]);

For further information, refer to the following documentation:

• https://www.mathworks.com/help/matlab/ref/readtable.html
• https://www.mathworks.com/help/matlab/ref/graph.htm
• https://www.mathworks.com/help/matlab/ref/graph.conncomp.html

Thanks

Chetan