Patched a bug in IsotopePatternGenerator #196
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Just to clarify, the test will still fail 50% of the time? Any ideas on a better method.
|
|
With the updates, the test method should not fail. |
|
For a better method, see e.g. http://www.sciencedirect.com/science/article/pii/S1044030507004308 Actually, calculation of exact isotope patterns is a non-trivial problem, as MW gets larger. Not sure if CDK also wants to cover the area of large molecules like proteins, though. |
|
Okay thanks for info and patch. |
johnmay
added a commit
that referenced
this pull request
Feb 16, 2016
Patched a bug in IsotopePatternGenerator
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I found this curious bug - running IsotopePatternGenerator on C20H30Fe2P2S4Cl4 returned sometimes 34 and sometimes 35 entries. Such non-deterministic behavior is not ideal for a library.
By tracking deeper I found the bug is caused by this: MolecularFormula.isotopes() returns an iterator of a Set, which is by definition unordered, and iterates the elements in random order (not sure how JRE manages this internally). The elements are therefore added to the growing isotope pattern in non-deterministic order. See https://github.com/cdk/cdk/blob/master/tool/formula/src/main/java/org/openscience/cdk/formula/IsotopePatternGenerator.java#L102
The method than does strange heuristics to decide whether to remove certain low-abundance isotopes. See
https://github.com/cdk/cdk/blob/master/tool/formula/src/main/java/org/openscience/cdk/formula/IsotopePatternGenerator.java#L224
However, whether the isotopes are low-abundant will depend on the order in which these are added to the calculation, which is random. Hence, the result is non-deterministic.
I added a junit test, which fails roughly in 50% of runs with CDK 1.5.12. I also added a patch for this bug, by reducing the threshold of what is considered low-abundance. Note: this does not fix the bug, only reduces its visibility (at the cost of increased computation time). Basically, I think the heuristics that is used is wrong, and the results of IsotopePatternGenerator are probably not 100% correct.