Biomedical Engineering Reference
In-Depth Information
parallel. In the end they need to be re-combined into a single table using
a cascade of Joiner nodes.
The computed features are all numeric or nominal, with the result that
the remaining parts of the workfl ow - building a predictive model and
applying it to unclassifi ed images - are straightforward (and therefore not
shown here; a similar example is given in Figure 6.1). All features originating
from the control samples are used to build the model, that is a decision tree
and all other segments are subsequently classifi ed using this model.
6.4.4 Next Generation Sequencing
The workfl ow described here (Figure 6.16) is an example of how to use
KNIME for Next Generation Sequencing (NGS) data analysis [7]. In
particular, it describes typical parts of a data analysis workfl ow with regard
to RNA-sequencing, DNA-sequencing, and ChIP-sequencing. It does not
try to show a complete or perfect workfl ow but rather to point out some
of the features of what can be done and explicit NGS relevant tools that
can be used. This and similar workfl ows are used by Institute Pasteur.
In general, this workfl ow reads-in FastQ formatted data, which is
cleaned and fi ltered, then aligned to a reference genome (hg19) using
bowtie [8] (upper part). The results are read and fi ltered and then written
Figure 6.16
A workfl ow for large-scale analysis of sequencing data
Search WWH ::
Custom Search