intersect file bam-bed and count read

[edeka]

Hi all!
I have RNA-seq data and I used STAR in order to align the reads.
I would like to use my bed file in order to have the only read that are mapped in the position reported in bed file and after I would like to know how many read are mapped in each interval.
For the first part I tried to use "bedtools intersect":
bedtools intersect -abam -a file.bam -b file.bed
I saw that the number of the read reported in the output file were less than the input file but It doesn't seem right. In fact, for example, in my bed file for chromosome the first interval begins with the position 335 and in my new output file there are reads mapped in position 80.

Thank you in advance and I am sorry for my probably bad English!
Best

[GenoMax]

The option you should look at is coverageBed.

As for the other observation if you are providing a set of BED intervals then only reads that fall in those intervals will be reported (so that number can be expected to be less than the input). It is possible that the read in question is mapped beginning at position 80 but depending on how long it is it must be extending into the first interval.

[SylvainL]

Since you did the mapping with STAR, it is possible a read will be really long once mapped to the genome if it is at an exon-exon junction... You can use the parameter -split in your bedtools command to report only the region your read is really mapping...

[edeka]

The option you should look at is coverageBed.

As for the other observation if you are providing a set of BED intervals then only reads that fall in those intervals will be reported (so that number can be expected to be less than the input). It is possible that the read in question is mapped beginning at position 80 but depending on how long it is it must be extending into the first interval.

Thank you for your reply. The read is about 20 nucleotides, so beedtools should discard this read...