Additional File 1:

Graph of MT families for HGU133A array. Animated GIF, 3D visualization of MT-families in the array.

Format: GIF Size: 8.5MB Download file

Additional File 2:

Examples of 3 families of probesets and transcripts. Screenshots from the applet. Big nodes signify probesets (green – positive detection call), small magenta ones – transcripts. The width of edges is proportional to the quantity of MT probes. Probes are marked with a name, annotation in Affymetrix or BioConductor and expression values. Presented are families associated mainly with PAX8, RUNX1/RPL22 and tubulins.

Format: TIFF Size: 423KB Download file

Additional File 3:

List of MT families for HGU133A array. The CSV file lists all the discovered MT probeset families along with their gene-level annotations according to Affymetrix and BioConductor.

Format: CSV Size: 286KB Download file

Additional File 4:

Applet for families exploration. http://bioinformatics.picr.man.ac.uk/adaptnet webcite. An applet for browsing graphs of MT-families in the HGU133A array. Big nodes represent HGU133A probesets: green ones have "Present" detection call, pink ones "Absent". They are labelled with Affymetfix and BioConductor annotations, detection call and expression value in the experiment. Small magenta nodes represent transcripts. There is a possibility to add Exon 1.0ST probesets (blue) to the graph. The width of edges is proportional to the number of matching probes. The applet is intended for online use – it is connected to an application server and ADAPT database.

Format: TXT Size: 1KB Download file

Additional File 5:

Spiking experiment, signal filtering 1. Scatter plot and correlation distribution, generated as in Figure 7, but filtered by average signal intensity. Low intensity: the 10% probesets with lowest mean signal. High intensity: the 10% probesets with highest mean signal. Low intensity spikes added to high targets. The plots in 5, 6, 7, 8 prove that with any sort of signal intensity filtering, the shift in correlation coefficient occurs.

Format: PDF Size: 101KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 6:

Spiking experiment, signal filtering 2. As 5, but high intensity spikes added to high targets.

Format: PDF Size: 100KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 7:

Spiking experiment, signal filtering 3. As 5, but high intensity spikes added to low targets.

Format: PDF Size: 100KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 8:

Spiking experiment, signal filtering 4. As 5, but low intensity spikes added to low targets.

Format: PDF Size: 100KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 9:

R code for correlation experiment. A simple experiment to consider the relationship between correlation cooefficient and variance using 10,000 randomly generated cases.

Format: R Size: 3KB Download file

Additional File 10:

Influence of specific number of spiked probes on correlation. Changes in Pearson correlation following spiking to simulate MT between probesets. Red plot – correlation before spiking, orange – 1 spiked probe per probeset, magenta – up to 3 probes, blue – up to 7 probes, green – all probes spiked. As in the case of real data – even a single probe may influence the distribution of correlation, however in that case there are no effects of biological similarity – that's why the effect exists, but is smallest for single probes.

Format: PDF Size: 10KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 11:

Distribution of the expression signal for MT and non-MT probesets after processing with RMA and MAS5. The plots (normalized distributions of summarized expression values) indicate a slight increase in the high signal values for MT probesets (blue) against non-MT probesets (green).

Format: PDF Size: 9KB Download file

This file can be viewed with: Adobe Acrobat Reader