Read atomic coordinates in a custom node.

Node Development
1

Hi there,

I have a quick question regarding creating a custom node that is able to read atomic coordinates, e. g. from an SdfCell object. I need the coordinates for some additional calculations inside the node. At the end an additional column with a number should appear, in which the result of the coordinate calculation is contained.

I took a look at the examples and some comments in the forum and the setup of the node seems pretty straight-forward. However I seem to cannot find how to extract the atomic coordinates out of the SdfCell in order to perform my calculations.

Any help greatly appreciated.

Best,

Endo

2

Could you parse the CTAB from the SdfCell to extract the coordinates? 

 

I expect one of the chemical engines could extract the coordinates. CDK may have a method like atom.getCoordinates() and you could iterate over the atoms int molecule?

 

You may also be able to do this with ChemAxon, I expect the API features you need would be covered under the API. 

Here is an example of getting X, Y and Z for each atom from a ChemAxon moleucule: https://gist.github.com/webbres/9d68e8f5d0498695ddb3 . 

 

Cheers

 

Sam

 

3

Thanks Sam,

I expected to find this functionality within the org.knime.chem.* or similar namespaces without the need to turn to third party packages. Since the molecules are drawn in the table in the UI dialog, there should be a method for getting the atoms.

I haven't tried your ChemAxon way, but I found out, that I can convert the molecules to CDK and then use CDKNodeUtils.calculateCoordinates(). However this seems a bit of an overhead and I was hoping that there is a more elegant way of doing this without an aditional library.

Any further suggestions would be much appreciated!

 

Cheers,

Endo

4

I don't believe any of the core KNIME provides anything more than the types. So you are unlikely to find any structure manipulation functionality there. (or am I mistaken?)

 

The rendering of the structures is drawn by one of the third party (partner/community) plugins you have installed. 

 

does calculateCoordinates generate the coordinates and overwrite those given?

5

Hi Sam,

yepp, that might very well be the case, that a third party plugin is rendering those and I looked  and found nothing, so I guess you are right.

It seems that calculateCoordinates generates new 2D coordinates but doesn't overwrite the old ones, if not forced. So it is not what I need anyway. With the CDK what is needed is indeed only this :

IAtomContainer cdkMol = SilentChemObjectBuilder.getInstance().newInstance(IAtomContainer.class);
        IteratingSDFReader reader = new IteratingSDFReader(new StringReader(referenceMoleculeCell.getStringValue()), SilentChemObjectBuilder.getInstance());  
            while (reader.hasNext())
            {
                cdkMol.add(reader.next());
            }

 

I'm currently looking into your ChemAxon example which seems equivalent. Thank you for providing me with those insights. You have been very helpful.

 

6

If you just need to access the co-ordinates, may be as easy in a java snippet to do something like this:

String[] sdfLines = sdfcell.split("\n");
int nAtoms = Integer.parseInt(sdfLines[3].split(" ")[0].trim()); //Check the 3 - it is the counts line in the sdf header block - it might be 2!

Double[] x = new Double[nAtoms];
Double[] y = new Double[nAtoms];
Double[] z = new Double[nAtoms];

for (int i = 4; i<nAtoms+4; i++) {
            //Loop through the atom lines to get the co-ords
            //Instead of putting all into arrays, you could use iteratively in your
            //Calculation if amenable
            x[i-4]=Double.parseDouble(sdfLines.substring(0,10).trim());
            y[i-4]=Double.parseDouble(sdfLines.substring(10,20).trim());
            z[i-4]=Double.parseDouble(sdfLines.substring(20,30).trim());
}

//Now do your stuff with the co-ords

Steve