regex in String Maniputation Node KNIME

KNIME Analytics Platform
#1

Hi All,

Hope you can help.

I have a large data set with a number of chemical formulas in a column of such data set. Some have square brackets along with charges (+ and -) that I want to remove. For example (totally made up formulas) it may look something like this:

[C6H6O27P7]13-
C6H19O27P7
[C6H6O23P7]2-
C36H38N2O6
C20H18O7
[C5H8N3O]+
[C5H8N3O2]2+
[C5H4N4O2]6+
C22H26O8
[C22H27O8]+

I cannot for the life of me figure out how to remove the brackets and charges in these locations. Would really welcome some help/advice!

The data set is large (a few hundred thousand compound) so manual removal is not an option!

Many thanks!

#2

Hi @rinawm

You could use the String Manipulation node with the configuration:

removeChars($formula$, "[]+-")

where formula is the name of the column containing the formulas from which you want to remove the brackets and charge signs.

Hope this helps!

2 Likes
#3

Are you sure that you want to mantain numbers outside square brackets?
How can you will distiguish which numbers were inside and which ones were outside the brackets?

#4

If you just want the content in the square brackets, then the Regex Extractor node (part of the Palladian node collection) will work if you use the expression:
[A-Z]\w+

These are the results I get:
regex

3 Likes
#5

Are you sure you don’t want to do a proper chemical operation on these structure representations - perhaps using RDKit or the Infocom wrapped ChemAxon libraries?

#6

Yes- I don’t need the charges. Key to this is the mol structure (on another column). Normally I’d keep them but there’s a bit of software that gets upset with the charges in the formulae!

#7

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.