Biomedical Engineering Reference
In-Depth Information
started blogging about their cheminformatics work. Willighagen started CML
Explained [65], Apodaca blogged on Depth First [66], Murray-Rust started the
CML blog [67] and PeterMR's blog [68], and Williams maintained the
ChemSpider blog [69]. Simply following each other's blogs turned out to be a
fairly effi cient way for the community to collaborate on shared interests. An
excellent example of this took place in September 2006 when Willighagen
suggested that the Open Source JSpecView applet could be used to view NMR
spectra in JCAMP-DX format [70], and this became immensely important to
sharing and manipulating raw data from UsefulChem [71], including the moni-
toring of reactions [72]. The ability to use JSpecView to overlay NMR spectra
was particularly useful for monitoring reactions [73] and measuring kinetics
when integrated with Excel VBA [74, 75].
Chemical Markup Language (CML) represented a promising way of openly
sharing chemical information. As we attempted to create CML really simple
syndication (RSS) feeds from our molecules blog, Willighagen and Murray-
Rust shared their expertise [76]. It may be better to characterize this type of
interaction as a metacollaboration since it did not involve project-specifi c
objectives so much as general ways of representing and manipulating chemical
information publicly. Lessons learned in this space would prove to be valua-
ble for quickly starting other chemistry open notebook projects, including
the conversion of laboratory notebook pages describing a Ugi synthesis into
CML [77] .
Willighagen proposed a method of introducing tags into blog posts to make
the molecules discussed machine readable [78]. This was experimented with
for the UsefulChem blog and provided a new means for potential collabora-
tors to fi nd information via the Chemical Blogspace, an aggregation service
for chemistry blogs. Further collaborations ensued. In June 2007 Guha created
a public Web service to generate a combinatorial list of all Ugi products result-
ing from lists of starting materials represented as SMILES [79]. Shattuck
created a Web service to deconvolute NMR spectra using JCAMP-DX fi les as
input [80], and in August 2007 an account was created on MyExperiment to
attempt to process and organize Web services related to the UsefulChem
project [81]. However, productive use of this system would have to await the
involvement of new collaborators after the creation of the ChemTaverna
project in 2010 [82].
In November 2007 enough data were being generated in the laboratory
notebook that it made sense to start abstracting Ugi reaction information into
a Google Spreadsheet to compile a CombiUgi Master Table [83]. Since each
record points to the corresponding laboratory notebook page, information is
not lost, but the abstraction allows for semantic querying of the data set.
Attempts were also made to convert the workfl ows into organized machine-
readable formats, involving a discussion between others (Williams, Willighagen,
and Murray-Rust) interested in overlapping objectives [84, 85]. In April 2008,
Guha created the fi rst version of a model to predict precipitation from metha-
nol based on molecular descriptors of the products [86].
Search WWH ::




Custom Search