Information Technology Reference
In-Depth Information
AffyExpressFilter1
and
GetPubMedAbstracts
are inserted, and accord-
ingly the
WriteTextFile
step at the end is replaced by a
WriteHtmlFile
step that stores the list of retrieved PubMed abstracts. However, naturally
also more comprehensive solutions are possible with this set of constraints:
the third example in the figure shows a solution similar to the second one,
where in addition the expression values are visualized before and after the
preprocessing step, and where finally an annotation table is created for the
top differentially expressed genes.
In fact, the essential general characteristics of micorarray analysis pipelines
are already covered by the defined domain constraints: When applying no
constraints at all, the search for synthesis solutions exceeds the 1,000,000-
solutions default when searching for solutions in search depths greater than
7, while with the domain constraints 1,779 solutions of length less or equal
than 7 are found. As already shown in the previous chapters for the other
three scenarios, especially for the loose programming of phylogenetic analysis
workflows in Section 3.3, the user can influence the synthesis results easily by
varying the constraints at workflow design time. Particularly effective is the
expression of concrete intents, for instance by explicitly demanding the use of
a specific data loading or annotation service. For example, the additional en-
forcement of
LoadSpikeInBenchmarkData
and
GetPubMedAbstracts
leaves
85 solutions that are found in search depth 7.
Search WWH ::
Custom Search