Jcup 3 (2012) Presentation: Lexichem, a new Era. By Ed Cannon
0 likes147 views
Lexichem is introducing a new version with improved performance in generating chemical names from structures and vice versa. The new version achieves 98.71% accuracy in "round tripping" structures compared to 60.02% for the previous version. It also offers faster processing speeds. New features include generating names that distinguish stereochemistry, support for ring templates, and the Lexichem Workbench desktop application for structure/name conversion and substructure searching. The performance enhancements and new tools provide chemists with more robust naming and structure conversion capabilities.
![Command Line Applications
Mol2Nam
• Convert a file of molecules to names
bicyclo[3.2.1]octane
bicyclo[3.2.1]octane
Nam2Mol
• Convert a file of names to molecules
Glycinate Translate
• Convert a file of chemical names into a different
グリシナート language
5](https://image.slidesharecdn.com/edwardcannonjcuprecorded-120523213736-phpapp01/85/Jcup-3-2012-Presentation-Lexichem-a-new-Era-By-Ed-Cannon-5-320.jpg)
![Octahydro-1H-4,7-epoxyisoindole Benzo[cd]indole
Ring
templates
5,6,6a,7-tetrahydro-4H-dibenzo[de,g]quinoline Yohimban
19](https://image.slidesharecdn.com/edwardcannonjcuprecorded-120523213736-phpapp01/85/Jcup-3-2012-Presentation-Lexichem-a-new-Era-By-Ed-Cannon-27-320.jpg)
Jcup 3 (2012) Presentation: Lexichem, a new Era. By Ed Cannon
- 1. Lexichem, a New Era
- 2. Overview • • • • • 2
- 3. Lexichem
- 4. Lexichem Nordefrin Supported Nomenclature Supported Languages (17) •IUPAC 79 / 93 / 2005 •English (American / British) •Chemical Abstract / CAS •German •Traditional •Japanese •MDL / Beilstein •Spanish •AutoNom •Swedish •OpenEye 4
- 5. Command Line Applications Mol2Nam • Convert a file of molecules to names bicyclo[3.2.1]octane bicyclo[3.2.1]octane Nam2Mol • Convert a file of names to molecules Glycinate Translate • Convert a file of chemical names into a different グリシナート language 5
- 6. Lexichem TK 6
- 7. Applications
- 8. Pipelines • Large scale conversion of structures to names and names to structures • Easy integration in workflows 8
- 9. Webservices • Lexichem Webservice available • Integration with 3rd party webservices • PubChem uses Lexichem Mol2Nam Molecule PUG Hits in s SOAP PubChem 9
- 10. Lexiparser • Automated extraction of structures and names from documents • Supported formats: - .txt - .docx - .html - .rtf - .doc - .pdf FUTURE 10
- 11. Extracting Structures from a Patent Patent URL Names extracted Generate Structures 11
- 12. Desktop Applications Electronic Laboratory Notebook NEW! Lexichem Workbench 12
- 14. Why a New Performance Metric ? • Ensures consistent improvement of Lexichem • Identify areas in need of development • Gold standard for all chemical nomenclature software 14
- 15. Round Tripping • Compare the initial Canonical and final structure after name generation Isomeric SMILES • Easy to calculate TP %RTCS = ×100 TotalStructures English IUPAC Name *E.O.Cannon. JCIM 2012, DOI: 10.1021/ci3000419 15
- 16. Results
- 17. Performance %RTCS 98.71 100.00 88.94 89.05 92.43 89.88 93.83 90.00 84.54 80.00 70.00 60.00 60.02 48.69 52.80 50.00 40.00 V2.0.2 30.00 V2.1.1 20.00 10.00 0.00 V2.1.1 Speed Maybridge MDDR V2.0.2 12000 NCI Wombat PubChem 10000 8000 Names s-1 6000 Molecules s-1 4000 2000 NCI Wombat PubChem 0 Maybridge MDDR NCI Wombat 17 PubChem
- 18. New Features
- 19. Nam2mol – New Features – 19
- 20. Nam2mol – New Features – – 18
- 21. Nam2mol – New Features – – – 18
- 22. Nam2mol – New Features – – – – H H H H H H H 18
- 23. Nam2mol – New Features – – – – – 18
- 24. Nam2mol – New Features – – – – – – 18
- 25. Nam2mol – New Features – – – – – – L-Arginine D-Arginine – 18
- 26. Nam2mol – New Features – – – – – – – – 18
- 27. Octahydro-1H-4,7-epoxyisoindole Benzo[cd]indole Ring templates 5,6,6a,7-tetrahydro-4H-dibenzo[de,g]quinoline Yohimban 19
- 29. Lexichem Workbench • • – – – – – – – 21
- 30. Main Window • Converts input SMILES string • Chemical information: or chemical name - Molecular weight • Visual display of the structure - SMILES - IUPAC name 22
- 31. Results • Results history • Original input on display 23
- 32. Results • Text options: - Copy selected cells - Save table - Display selection 24
- 33. Results • Display options: - Save - Copy - Print 25
- 34. Substructure Search • Options: - Functional group from list - Custom SMILES/SMARTS pattern - Custom name 26
- 35. Conclusions • • • 27
- 36. • • • • • • 28
Editor's Notes
- #3: Start by talking about some real world applications of Lexichem, what you can do with it and who is using it.then move on to talk about Lexichem, TK v2.11….released Metric to assess how well Lexichem is performingNew features from v2.0.2Finish off with GUI
- #4: Now lets talk about the main purpose of my were here, Lexichem.
- #5: Lexichem is OpenEyes chemical nomenclature software, you can:->convert names to molecules->molecules to names->translate names from one language to anotherLexichem comes in two flavours
- #6: Standalone applications run from command line.
- #7: For those who want to program and use Lexichem:Lexichem TK is written in C++ and Swig (Simplified Wrapper and Interface Generator) wrapped to python, Java and C#.
- #8: So what can Lexichem do, other than help you buy heroin?
- #9: Keith Taylor showed yesterday a Pipeline pilot workflow with a node for converting structures to namesWorkflow integration with Pipeline Pilot.Node use 1 of Lexichem’s functions-> mol2nam, nam2mol, translateMatt Stahl working on nodes using OpenEye Software.Lexichem node highlighted in square ->convert structures to names.
- #10: Craig Bruce hired recently, been developing Webservices for OE, one being LexichemUse Lexichem prior / post processing to Webservice.PUG (Power User Gateway)Search for structures with the names found in across PubChem.
- #11: Chemical name extraction from patents/documents and structure generation (Lexiparser uses Lexichem)->Uses Lexichem after it’s extracted chemical names to convert them to molecules.
- #13: Lexichem is the engine beneath the hood, when a user draws a structure a call to Lexichem is made which generates a name which can be rendered.Alternatively you can import chemical names convert them to molecules and visualize the image.
- #14: Purpose of metric-> ensure we are not regressing but improving Lexichem-> identify areas / features in need of improvement-> gold standard which other companies can then compare chemical nomenclature software
- #15: ->Concept: start pt, and an end pt after some processing, then compare start pt to end pt+ve: quick to calculate, gives one figure value of how accurate Lexichem is on dataset->Paper acceptedAdv SMILEs: human readable, less verbose than inchi, tautomer support
- #16: ->Concept: start pt, and an end pt after some processing, then compare start pt to end pt+ve: quick to calculate, gives one figure value of how accurate Lexichem is on dataset->Paper acceptedAdv SMILEs: human readable, less verbose than inchi, tautomer support
- #17: The concept of a benchmark is good, but do we have good results using it?
- #18: Whilst these results are good, is Lexichem feasible on a large scale?Seen Lexichem performs well and is feasible on large datasets, now lets look at what features been added.----- Meeting Notes (3/28/12 11:38) -----Mol2Nam -> canonicalize atoms & bonds,identify atom types, identify ring systems and size, bridges, locants and positions,identify stereo / walk the graph
- #19: So what have we added since v2.0.2?Our main drive has been looking nam2mol features (as they’re not quite as well supported as mol2nam), in particular the ability to generate molecules for large ring systems. (one of Lexichem’s weaker points in the previous releases)
- #20: Von Baeyer -> polyalicyclic ring systems-> previously only bicyclic supported.-> working hard on augmenting natural productsBeta-carotene in carrots -> Provitamin A carotenoid
- #21: Von Baeyer -> polyalicyclic ring systems-> previously only bicyclic supported.-> working hard on augmenting natural productsBeta-carotene in carrots -> Provitamin A carotenoid
- #22: Von Baeyer -> polyalicyclic ring systems-> previously only bicyclic supported.-> working hard on augmenting natural productsBeta-carotene in carrots -> Provitamin A carotenoid
- #23: Von Baeyer -> polyalicyclic ring systems-> previously only bicyclic supported.-> working hard on augmenting natural productsBeta-carotene in carrots -> Provitamin A carotenoid
- #24: Von Baeyer -> polyalicyclic ring systems-> previously only bicyclic supported.-> working hard on augmenting natural productsBeta-carotene in carrots -> Provitamin A carotenoid
- #25: Von Baeyer -> polyalicyclic ring systems-> previously only bicyclic supported.-> working hard on augmenting natural productsBeta-carotene in carrots -> Provitamin A carotenoid
- #26: Von Baeyer -> polyalicyclic ring systems-> previously only bicyclic supported.-> working hard on augmenting natural productsBeta-carotene in carrots -> Provitamin A carotenoid
- #27: Von Baeyer -> polyalicyclic ring systems-> previously only bicyclic supported.-> working hard on augmenting natural productsBeta-carotene in carrots -> Provitamin A carotenoid
- #28: Ring templates for conversion of molecules to names.Mainly bridge and fused ring templates have been added.
- #29: Primary goal of the GUI was to:-> Lower the bar to use Lexichem’s functionality (for people not keen on using an API to program against, or using the command line)
- #31: ->Primarily modeled on the command line tools, but provided numerous additional features2 Menus: Open a molecular fileImport name, SMILES, set default options, clear the view.
- #35: Filter the results
- #36: Story about Lexichem, you know you want it, then please feel free to contact usFuture work: continue to work on fused polycyclic ring systems, natural products