Tutorials

2.1 From the picture to the raw data

If we had the images, we could quantify the amount of each of the fluorescent probes in every spot, remove results for problematic spots. This could be done with one of the many software available like Scanarray, GenePix. It will also be necessary to normalize the arrays. We can use SNOMAD for this task, for example.

This step is out of the scope of this tutorial: we will use public data sets already analysed by other groups.

2.2 From the raw data to the gene expression patterns

We have to merge the results of all the experiments we want to use into one file. If our experiments are in a Data Base, we will have to use the available tool for extracting the results. If we have the results in flat file, we could use a text editor or a SpreadSheet.

Next, we have to preprocess the patterns. This step includes the transformation of the scale, the handling of missing values, the removing of flat gene expression patterns and the standardization of the remaining patterns. Those steps are optional and depends on the actual data set.

To know more about preprocessing, see the tutorial on gene expression pattern.

2.3 Clustering of gene expression patterns

This step will group the gene expression patterns. We can use any of the available methods and compare the results.

2.4 Analysis of the results: data mining

The first analysis of the result of a clustering is the comparison of the functions of the genes that cluster together and the differences with the function of the rest of the genes.