Why is there a big difference between the outputs of `geneBody_coverage` of `RSeQC` when using different bed files?

Question

[Dan]

Hi

I run geneBody_coverage.py of RSeQC (https://rseqc.sourceforge.net/#contact) using different bed files and the same batch of bam files:

geneBody_coverage.py -r long_short_genes/changeformat/hg19.RefSeq.union.bed \
-i split_bam/short_genes/bams/ \
-o geneBody_coverage/short_genes/RefSeq

The output is very different.

What is difference between union.bed files and other bed files which I downloaded from the following links:

https://sourceforge.net/projects/rseqc/files/BED/Human_Homo_sapiens_merge_transcripts/

https://sourceforge.net/projects/rseqc/files/BED/Human_Homo_sapiens/

Which bed file should I use?

Thanks

[GenoMax]

If you scroll down the pages you linked above there are README contents displayed that explain how the "union" files are made.

[Dan]

Thanks. I still don't understand why the output is so different: the output of hg19_Ensembl_gene.bed the output of hg19.Ensembl.union.bed

[Dan]

I also run geneBody_coverage.py using different .bed files, hg19_Ensembl_gene.bed and hg19_RefSeq.bed and the same batch of bam files, but the outputs are very different:

geneBody_coverage.py -r hg19_Ensembl_gene.bed \
-i split_bam/short_genes/bams/ \
-o geneBody_coverage/short_genes/Ensembl_gene

geneBody_coverage.py -r hg19_RefSeq.bed \
-i split_bam/short_genes/bams/ \
-o geneBody_coverage/short_genes/RefSeq

I am very confused about why they are so different.

Which bed file should I use?

Thanks