Please wait while the transcript is being prepared...
0:00
Hello, my name is Johan A. Westerhuis and I'm from the University of Amsterdam.
I'm going to give you an overview of Data Fusion Methods,
and I will show some examples of fusing Metabolomics and Transcriptomics Data.
0:13
On this next slide,
you can see the content of my talk.
I will first introduce the difference between data integration and data fusion.
and we'll discuss some of the goals of data fusion,
and after that I will show some of the methods
used and some complex issues in data fusion.
0:31
In this next slide,
you can see an example of data integration by multiple sources of omics data,
and also ontologies and databases that are combined.
On this map, you can see an overview of the glycolysis of some data
and where we have measured metabolomics data, RNA-seq,
miRNA, proteomics and even more sources of data.
Now, because this is the glycolysis,
we will have to know information about each gene and
where it's active in this map.
And therefore, we need the ontologies,
and we might even use some of the databases
for example, consisting of mRNA and miRNA associations.
So, together the whole overview of all of these different data sources,
is what I call data integration.
1:14
On this next slide,
I'm going into the data fusion methods.
Data fusion methods are sort of statistical methods which can combine different datasets,
measured on the same samples or on the same individuals.
And the most important sort of issue, is that measurements on
the same samples should appear in all of the datasets simultaneously.
Here, you see an example
of fusing Metabolomics data measured on GCMS, LCMS and NMR.
And here you see, indicated by this line,
that samples on the same row in each of these sets,
that they belong together;
so, to the same individual or to the same sample.
And there are other examples of fusing metabolomics with gene expression and proteomics,
or even fusing different compartments.
And here we see an example of kidney, plasma,
and liver tissues all measured from the same individual.
And then, sort of the same individual,
has the same level of information that we want to combine.
So, what are the goals of data fusion?
