Skip to first unread message
Rajan Joshi
Jan 1, 2017, 2:29:42 PM1/1/17
to MEME Suite Q&A
If I run glam2 on the server with the default settings: glam2 -o set851Out p set851.fasta I get the score of 4612.64. If I am to run the same sequences shuffled 5 times using the Emboss Shuffled (as recommended in the Glam2 manual to improve the score) my score jumps to 19266.8. The score goes up because there are five times the number of sequences so the score should be greater. After reading some papers and the manual, I understand the glam2 output score of each motif represents the addition of individual scores added for each sequence matching to the motif.
My question is how do I interpret the output scores of glam2 results? What score makes a motif significant? Because the larger number of sequences you provide the larger your score gets. In MEME the significance is determined by the E-value. The lower the E-value the more significant the motif correct? A motif with an E-value above .05 is not significant. But how so for glam2 results? Thank you for the help or suggestions.
CharlesEGrant
Jan 4, 2017, 2:58:41 PM1/4/17
to MEME Suite Q&A
Glam2 doesn’t provide a built in statistical significance estimate for the motifs it discovers. The glam2 score provides a relative measure of the motif quality within a single data set, but it’s not calibrated, so you can’t directly compare scores between data sets.
The glam2 tutorial discusses a couple of methods using multiple runs of Glam2 which can be used to estimate the statistical significance of motifs. See the section at the bottom of the page titled "Significance of glam2 motifs”:
The glam2 tutorial discusses a couple of methods using multiple runs of Glam2 which can be used to estimate the statistical significance of motifs. See the section at the bottom of the page titled "Significance of glam2 motifs”:
Reply all
Reply to author
Forward
0 new messages