How much overlap to consider when annotating a region

Question

[Saad Khan]

Hi I have chip and Dna methylation data in 300bp windows. I am trying to annotated these windows as falling into different kinds of regions e.g. intergenic,exon,intron,utr,promoter etc. I have already fetched this information from gtf file using the method described here using bedtools. Now I want to annotate my windows as either falling into either of these regions. I am confused about a few things and would be glad if someone could help me out.

1. What percentage of overlap should I consider in order to unambiguously divide my tiles into either of the annotation.
2. If I don't consider overlap and only consider regions to be belonging to certain kind of annotation then what parameters can be best used to do it.

Answer 1

[PoGibas]

Simple answer would be: "Are you ready to try some R?!" :-)

See my answer in Overlap between 2 sets of genomic regions of differing size - use GenometriCorr, it will produce reliable results and nice visualisations.