Extract unique features from BAM/BED

[rxzlmn]

Hi,

I'm trying to extract 'unique' features from either BED or BAM. I have small RNAseq data that I want to use to identify novel short RNAs transcribed from a specific genomic region. Those short RNAs overlap, but I don't want to assemble them into a larger transcript, but rather extract all 'unique' short RNAs. Each of those overlapping short RNAs is also covered by many reads.

Basically I want to perform something like mergeBed, but forcing a 100% similarity. Is there a tool available that would allow me to do that?

[GenoMax]

How about using intersectbed (http://bedtools.readthedocs.org/en/l...e-usage.html)?

[dariober]

Basically I want to perform something like mergeBed, but forcing a 100% similarity. Is there a tool available that would allow me to do that?

I'm not sure I completely understand your task, but if you want to extract all the unique intervals in a bed file you could use:

cut -f1-3 reads.bed \
| sort -k1,1 -k2,2n -k3,3n \
| uniq -c \
| awk 'BEGIN {OFS="\t"} {print $2, $3, $4, $1}' > unique.txt

The last column of unique.txt is the count of each interval.

Dario