GV Exercise.8

Find genes which expression correlates (anti-) with one given marker

[ Main_Page | Genevestigator_training | Analyze_public_microarray_data_using_Genevestigator | GV Exercise.7 ]
last edit: October 31, 2014

---

Aim

Find probes that show expression profiles correlated with that of a given probe (gene). Probes showing similar expression profiles across conditions could be co-regulated and therefore be part of a shared biological process. The principle of guilt by association is often used to discover genes related by some biological function or process.

Find other genes similar in expression to one of the ABA-1h probe from the former exercise (ASG4 259432_at aka AT1G01520)

You can rename items in the list using the Gene Label button

In the context of ALL Ath experiments

• Use the ABA signature list and set the Sample selection back to Ath.All and uncheck all probes above 259432_at to have this probe first checked in the list

try it first

• note that hovering onto the probe shows more information

try it first

• start the coexpression tool and search for 25 correlated probes in the perturbation context

try it first

• run the tool for positive correlations

try it first

the correlation values are quite low due to the very broad context context

In the context of the AT-00110 experiment

To improve correlation results, we need to focus on conditions that are more related to each other. In this example, we choose the easy way by taking the experiment than was used to build the signature, In real cases, you will chose samples based on other GV result as seen in former exercises.

• switch to the AT-00110 sample group and run the tool for positive correlations

try it first

• run the tool for negative correlations

try it first

Continue the analysis outside of Genevestigator

• export this data to table in order to use it elsewhere

positive correlations

• run the tool with anti-correlations to discover inverse profiles and export the data in excel format

negative correlations

• perform functional enrichment of the two lists in Plaza, Plant GSEA, or any tools that accepts gene lists (and/or correlation coefficients). To do so, please proceed as shown in GV_Exercise.4#Perform_GSEA_with_the_Plant_GSEA_web-portal.

There seems to be a possibility to transform Pearson correlation coefficients into p-values . The problem is to define the sample size - all genome! or sample size itself (25)

Download the exercise files

Try it by yourself before expanding on the right!

• download ex8.gv4 and open it from within genevestigator File Load Workspace link
• download the signature worksheet link
• download the ATG4 positive correlations link
• download the ATG4 negative correlations link

---

References:

---

[ Main_Page | Genevestigator_training | Analyze_public_microarray_data_using_Genevestigator | GV Exercise.7 ]