Crowdsourced Curation of Chemistry Data. How Bad is Online Chemistry Data? Antony Williams Wolfram Summit, September 2010

A Pragmatic Vision “Build a Structure Centric Community”  Integrate chemistry across the internet based on “chemical structure”  A “structure-based hub” to information and data  Let chemists contribute their own data  Allow the community to curate/correct data

We Answer Questions for Chemists  Questions a chemist might ask…  What is the melting point of n-heptanol?  What is the chemical structure of Xanax?  Chemically, what is phenolphthalein?  What are the stereocenters of cholesterol?  Where can I find publications about xylene?  What are the different trade names for Aspirin?  What is the NMR spectrum of Benzoic Acid?  What are the safety handling issues for toluene?

Search for a Chemical…by name

Available Information…  Linked to vendors, safety data, toxicity, metabolism

Available Information….

Search for chemicals

ChemSpider Today  24.8 million structures  400 data sources  Grows daily  Community annotation and curation  We curate, edit, change, enhance data daily

Linked Data on the Web

Three Years of Experience  Internet-based chemistry is a mess!  Most public compound databases on the web are contaminated. Including ours!  The annotation/curation of data online is difficult  Most database hosts are non-responsive to feedback – “We are a host/repository of data”  Who cares?

Where is chemistry online?  Encyclopedic articles (Wikipedia)  Chemical vendor databases  Metabolic pathway databases  Property databases  Patents with chemical structures  Drug Discovery data  Scientific publications  Compound aggregators  Blogs/Wikis and Open Notebook Science

What is the Structure of Vitamin K?

MeSH – Medical Subject Headings  A lipid cofactor that is required for normal blood clotting. Several forms of vitamin K have been identified: VITAMIN K 1 (phytomenadione) derived from plants, VITAMIN K 2 (menaquinone) from bacteria, and synthetic naphthoquinone provitamins, VITAMIN K 3 (menadione). Vitamin K 3 provitamins, after being alkylated in vivo, exhibit the antifibrinolytic activity of vitamin K. Green leafy vegetables, liver, cheese, butter, and egg yolk are good sources of vitamin K

What is the Structure of Vitamin K1?

Chemical Abstracts “Common Chemistry” Database

Wikipedia

Incorrect Structures

Wow!

Lack of Stereochemistry

Does stereochemistry matter?  Distaval, Talimol, Nibrol, Sedimide, Quietoplex, Contergan, Neurosedyn, Softenon, Thalidomide

PubChem

“2-methyl-3-(3,7,11,15-tetramethylhexadec-2- enyl)naphthalene-1,4-dione”  Variants of systematic names on PubChem  2-methyl-3-[(E,7R,11R)-3,7,11,15-tetramethyl  2-methyl-3-[(E,7S,11R)-3,7,11,15-tetramethyl  2-methyl-3-[(E,7R,11S)-3,7,11,15-tetramethyl  2-methyl-3-[(E,7S,11S)-3,7,11,15-tetramethyl  2-methyl-3-[(E,11S)-3,7,11,15-tetramethyl  2-methyl-3-[(E)-3,7,11,15-tetramethyl  2-methyl-3-(3,7,11,15-tetramethyl  2-methyl-3-[(E)-3,7,11,15-tetramethyl

ChEBI – Manual Curation

What’s Methane?

What ELSE is Methane???

The EXPERTS must get it right?!

Wikipedia, C&E News, PubChem C&E News (from ACS)

Internet-Based Chemistry is a Mess  Algorithms can get you so far  Human curation is necessary  Only the crowds can help with big data… ChemSpider is approaching 25 million compounds

Search “Vitamin H”

“Curate” Identifiers

 General curation activities  Remove incorrect names  Correct spellings  Add multilingual names  Add alternative names  In 3 years over 1 million structure-identifier relationships have been validated – robotically and manually  130 people have participated in validation or annotation. “Crowds” can be quite small!

Crowdsourced “Annotations”  Registered Users can add  Descriptions/Syntheses/Commentaries  Links to articles, blogs, wikis etc  Add spectral data  Add photos  Add MP3 files  Add Videos

Data Validation – Not Vitamin K1

Data Validation – Not Beclamethasone Dipropionate DailyMed Article

Data Validation …NOT Cholesterol

Data Validation – ONE Cymarin Question Quality in Big Databases

First request to Database Hosts!  Every public compound database host should add ONE feature – “Leave Comments”

Second request to Database Hosts! Show Comments

Always Question Online Chemistry

Thank you Twitter: ChemConnector Blog: Personal Blog: SLIDES: