Finding significant motif binding sites in clusters

---

 

Starting Out

To begin working with Sequences and Motifs, go to Motifs menu > Open Motif File. This will pop up a dialog box allowing you to select a GeneXPress Motif file (.gxm) containing PSSMs. Choose the file you would like to analyze and click Open.

Motif information

The top left corner of the screen indicates the currently selected cluster. The promoter region of each gene in the cluster is displayed in the large bottom section of the screen. Each motif "hit" is lit up. You can explore the genomic regions by scrolling around this pane. Finally, for each motif, a chart is displayed plotting for each cutoff score the ratio of genes that have this motif from the genes in the cluster vs. the ratio of genes that have this motif from genes outside the currently selected cluster. The vertical black line indicates the most significant cut-off score (by p-value.) This score from Motif #1 also determines the threshhold for lighting up motif "hits."

Analyzing current cluster for motif enrichment

When a motif is loaded, statistics are computed to obtain the significance of the motif on each cluster. As you switch between clusters (i.e. from BirdsEye View), and go back to the sequence screen, you will see the chart mentioned above plotting the hit ratios and most significant cut-offs for those clusters. If you'd like to analyze all your motifs on all clusters (say to find clusters significantly enriched for some motif), go to Analyze > Motifs, select All Clusters and the largest p-value you would consider interesting and click the Analyze button. The statistical computations and analysis may take a few minutes depending on the number of motifs and clusters you have, but only the first time they are performed, the results are saved afterwards and can be accessed throughout your GeneXPress session.

Results

The product of these computations is a table, (which can be sorted by pvalue by clicking on that column name OR sorted in reverse order by doing SHIFT+click) Each record is an instance of a significant set of occurences of one of the motifs in the gxm file in one of the clusters in your test results. Clicking (or scrolling to) a row in the table will cause the current cluster to change to that cluster, the motif screen to change Motif #1 to that motif and the threshhold score to be set to the best cut-off for that motif on that cluster. You can then examine that motif on that cluster as described above.

Addtitional Tips

To gain a more global idea of the location of motif hits in the promoter regions of the genes in this cluster (perhaps to notice spatial patterns), try setting the pixel width and height to smaller values (i.e width=2 height=1). You can vary the zoom factors throughout your session with GeneXPress (the sequence letters appear starting with 8,8)

 

 

Thank you for taking the GeneXPress tutorial. Click to continue:

Previous:	Find Significant Functional Groups in Your Clusters
Home:	Tutorials Home