lifs-tools/goslin

Goslin is the Grammar on succinct lipid nomenclature.

goslin

Goslin is the Grammar on succinct lipid nomenclature.

Goslin defines multiple grammers compatible with ANTLRv4 for different sources of shorthand lipid nomenclature. This allows to generate parsers based on the defined grammars, which provide immediate feedback whether a processed lipid shorthand notation string is compliant with a particular grammar, or not.

Overview of Goslin and Tutorials

Goslin 2.0 supports the updated lipid shorthand nomenclature with new structural levels.

Citing Goslin

If you use Goslin or any of the specific implementations in your work, we kindly ask you to cite the original publication:

• D. Kopczynski et al., Goslin - A Grammar of Succinct Lipid Nomenclature, Analytical Chemistry, June 26th, 2020. doi:10.1021/acs.analchem.0c01690

If you are using any of the new features of Goslin 2.0, please cite the following, updated Goslin 2.0 publication:

• D. Kopczynski et al., Goslin 2.0 Implements the Recent Lipid Shorthand Nomenclature for MS-Derived Lipid Structures, Analytical Chemistry, April 11th, 2022. doi:10.1021/acs.analchem.1c05430

References

• D. Kopczynski, N. Hoffmann et al., Analytical Chemistry, April 11th, 2022
• D. Kopczynski, N. Hoffmann et al., Analytical Chemistry, June 26th, 2020
• D. Kopczynski, N. Hoffmann et al., Biorxiv, April 20th, 2020 (Preprint)

Related Projects

• This project
• C++ implementation
• C# implementation
• Java implementation
• Python implementation
• R implementation
• Webapplication and REST API

Test data

1. testfiles/lipidmaps-names-Feb-10-2020.tsv - generated from LipidMAPS LMSDB export on Feb. 10th, 2020. Filtered all entries without an abbreviation.
2. testfiles/swisslipids-names-Feb-10-2020.tsv - generated from Swiss Lipids (lipids table) export on Feb. 10th ,2020.

Short samples of lipid names used for testing of the implementations are available from the testfiles directory.

License

The Goslin grammars are licensed under the terms of the MIT license (see LICENSE).

goslin's Issues

MG, DG, TG in Goslin parser

Add support for "MG", "DG", "TG" in the old Goslin parser.

Sterols with old nomenclature

Output is "ST 28:2;1" instead of "ST 28:2;O".

Split grammars and import common blocks

ANTLR v4 allows import of grammars:

https://github.com/antlr/antlr4/blob/master/doc/grammars.md#grammar-imports

Ether lipid "O-" not recognized

Describe the bug
Ether lipids are misclassified as Di-acyl-species

To Reproduce
Steps to reproduce the behavior:

1. Go to Goslin webapplication v. 1.1.4
2. Enter "PC O-18:1_18:1"
3. click validate
4. See result as screen shot: not working . sn-1 ether connection is not recognized but classified as "Ester". Nevertheless, the lipid is named as "PC O-18:1-18:1"

Expected behavior
The same works when using "PC P-18:1/18:1".
See result as screen shot: working.
Else: If the lipid name is not correct, there should be an error. But the O- is transferred without error.

Screenshots
If applicable, add screenshots to help explain your problem.

not working:

working:

Additional context
Thank you for your work. We plan to use Goslin starting with version 2.0.

Problems parsing complex HMDB names

Some HMDB lipid names cannot be parsed.
I have implemented some patches in @gbaquer/LipidParser. Happy to incorporate into goslin and test. Let me know.

Issue 1: Functional groups

• PG(20:4(8Z,11Z,14Z,17Z)-2OH(5S,6R)/22:6(4Z,7Z,10Z,13Z,16Z,19Z))
• PC(20:5(7Z,9Z,11E,13E,17Z)-3OH(5,6,15)/14:0)
• PS(18:2(10E,12Z)+=O(9)/18:1(9Z))

Solution:
Remove with REGEX
x=re.sub('(-\w+\(\d+[ A-Za-z,\d]*\))|\+=\w\(\d+\)','',x)

Issue 2: Specific FA's

• PA(P-16:0/LTE4)
• PE(PGJ2/20:5(5Z,8Z,11Z,14Z,17Z))
• PE(22:0/5-iso PGF2VI)

Solution:
Replace by the corresponding nomenclature. For example: PGJ2 -> 20:5
for i in range(fa.shape[0]): x=re.sub(fa.name[i],fa.repname[i],x)

Update LIPID MAPS Subclass mappings for Glycosphingolipids and Globosides

Is your feature request related to a problem? Please describe.
At the moment, Glycosphingolipids and Globosides are not directly mapped to a Lipid Maps class or subclass.

Describe the solution you'd like
The should be mapped to their equivalent Lipid Maps class or subclass.

Finalize modifications once updated nomenclature has been published

Currently, only the hydroxyl modification is handled and accounted for during the parsing.
Other modifications are currently parsed into the datamodel, but are not handled.

Implement a lipid report function

Implement a lipid report function for all the implementations as it is in rGoslin.

Generate SMILES from lipid building blocks

Parsing Ubiquinone or Coenzyme

Hi,

I am currently using rgoslin release version 1.1.2

I was trying to use it to parse a Ubiquinone, namely Coenzyme Q10.

Unfortunately, I do not know what name should I use as an input.

• CoQ 10
• Coenzyme Q10
• Ubiquinone-10

library("rgoslin")

isValidLipidName("CoQ 10")
#> Warning in rcpp_is_valid_lipid_name(lipidName): Parsing of lipid name 'CoQ 10'
#> caused an exception: Lipid not found
#> [1] FALSE
isValidLipidName("Coenzyme Q10")
#> Warning in rcpp_is_valid_lipid_name(lipidName): Parsing of lipid name 'Coenzyme
#> Q10' caused an exception: Lipid not found
#> [1] FALSE
isValidLipidName("Ubiquinone-10")
#> Warning in rcpp_is_valid_lipid_name(lipidName): Parsing of lipid name
#> 'Ubiquinone-10' caused an exception: Lipid not found
#> [1] FALSE
Created on 2021-09-02 by the reprex package (v2.0.1)

Kindly advise.

Thank you.

Update LIPID MAPS grammar for updated shorthand notation

Is your feature request related to a problem? Please describe.
LIPID MAPS has recently updated their shorthand nomenclature / abbreviation following this publication:
https://www.jlr.org/content/early/2020/10/09/jlr.S120001025.full.pdf

However, there are still some cases, where the new rules (removal of unnecessary parentheses) are not displayed in the common name, but only on the abbreviation levels:
https://www.lipidmaps.org/data/LMSDRecord.php?LMID=LMGL02010003

Describe the solution you'd like
Goslin should be able to parse the updated names and should also respect the updated hierarchy introduced in the paper linked to above.

Describe alternatives you've considered
None

Additional context
None

Add support for MS-DIAL 4 nomenclature

The MS-DIAL 4 nomenclature seems to be reasonably close to the existing nomenclatures to be able to add it with little effort:
http://prime.psc.riken.jp/compms/msdial/lipidnomenclature.html

ConstraintViolationException for omega-linoleoyloxy-Cer

The error is: Number of specified fatty acyl chains (2) not allowed for lipid class 'omega-linoleoyloxy-Cer' (having 0 fatty aycl chains).

Handle LipidMaps sphingomyelin names

Currently, names like "SM 21:0;2O/21:1" are not successfully parsed.

Introduce grammar for fatty acids

Add Lipidclass nomenclature from ALEX123

For LUX score template SMILES, the naming conventions and list of lipid classes of Pauling, Hermansson, et al. PLoS One 2017 (http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0188394) was used.
Some Lipidclass names are not used in Lipidmaps or Swiss lipids and would be great to have as an additional alias or complete addition, such as:

SQDG ~ Sulfoquinovosyl diacylglycerol (https://www.lipidmaps.org/data/LMSDRecord.php?LMID=LMGL05010004)
aPG = n-acylated PG (no entry in DBs, fatty acid at the head group)
aPE = NAPE ~ additional name
DMPE = PE-NMe2 ~ additional name
MMPE = PE-NMe ~ additional name
PEt = PEth = PEtOH ~ Phosphatidylethanol without the amine part

HeadGroup Formula for Ceramides is missing two Oxygen

The sum formula calculation for Ceramides is incorrect, since Ceramides have two additional Oxygens, which are not included in the head group definition of Goslin:

SP | Ceramides [SP02] | [2] | 2 | H | 1.0078 | [Cer, Ceramide]

This is in the following line of the table

goslin/lipid-list.csv

Line 64 in 89dc3ae

Cer,SP,Ceramides [SP02],2,2,H,Ceramide,,,,,

To Reproduce

1. Go to https://apps.lifs.isas.de/goslin
2. Enter Cer 30:1 and Cer 18:1/16:0 into the form
3. Click on submit
4. Check the sum formulas and corresponding m/z values.
5. The sum formula is C30H59NO and the neutral exact mass is 449.4597

Expected behavior
The sum formula should contain two more Oxygens, it should read HO2. Thus, the full sum formula should return C30H59NO3 in case of Cer 30:1

Translation system for trivial names

Lipids with trivial names such as EPA should get a more clever system for translating into the internal data structures.

Add ACer to Goslin grammars

Is your feature request related to a problem? Please describe.
ACer (Acyl Ceramides) are currently not supported in the Goslin Grammars.

Describe the solution you'd like
ACers should be parseable and mapped similarly to other Ceramides.

ACers are currently not available in LIPID MAPS apparently.
SwissLipids has some examples: https://www.swisslipids.org/#/search/Acylceramide

The raise_error flag is not suppressing all exceptions

The raise_error flag is not suppressing all exceptions in all the implementations.

Dependency Dashboard

This issue lists Renovate updates and detected dependencies. Read the Dependency Dashboard docs to learn more.

Open

These updates have all been created already. Click a checkbox below to force a retry/rebase of any.

• Update gradle/gradle-build-action action to v2 [SECURITY]
• Update dependency org.antlr:antlr4 to v4.13.1
• Update actions/checkout action to v4
• Update actions/setup-java action to v4
• Click on this checkbox to rebase all open PRs at once

Detected dependencies

github-actions

.github/workflows/check-grammars.yml

• actions/checkout v2
• actions/setup-java v2
• gradle/gradle-build-action v1.5.0

gradle

build.gradle

• org.antlr:antlr4 4.9.2

---

• Check this box to trigger a request for Renovate to run again on this repository

Lipids with "dimethylarsinoyl" cause an exception

Name 'MMAs' not registered in functional group list