[ensembl-dev] Predicting Motif Score Change

[Rashid Mamunur]

Hi everyone,

I am trying to predict the predict the effect of a variation on the motif binding affinity. I have extracted the values via VEP.
However to understand the mechanism better, I am trying to compare VEP information with manually (Ensembl FuncGen API) 
extracted information.

Example : 
1	40505680	T	C
For the above genomic location I try to extract motif features and original binding affinity and I get following information :
 
Srf:MA0083.1 -> GACCATCTATGG | 40505673 | 40505684 | 1 | 12 | MA0083.1 | affinity:	0.888415882885811
Srf:MA0083.1 -> GACCATAGATGG | 40505675 | 40505686 | -1 | 12 | MA0083.1 | affinity:	0.888745476322399

This is an example of same motif in forward and reverse strand. My question is 
- if I want to calculate effect of variation, which bases in the above two sequences do I change ??

For the same location predicted with VEP, I have the following result  :  

	MOTIF_POS=7;MOTIF_NAME=Jaspar_Matrix_Srf:MA0083.1;HIGH_INF_POS=Y;MOTIF_SCORE_CHANGE=-0.092
	MOTIF_POS=8;MOTIF_NAME=Jaspar_Matrix_Srf:MA0083.1;HIGH_INF_POS=Y;MOTIF_SCORE_CHANGE=-0.092

As VEP output does not give any strand information, I am struggling to map the VEP outputs to the motifs extracted via FuncGen;
- which positions in the original motif sequence does the MOTIF_POS values corresponds to ??

Apology in advance for such a long question. 

Thanks in advance.

Mamun
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ensembl.org/pipermail/dev_ensembl.org/attachments/20140220/79b15176/attachment.html>