@Pierre Thanks for your reply, can I also extract one mpileup per read with a filtering region II know try it like this:

while read LINE
do
    chr=$( printf "$LINE" | cut -f1 )
    start=$( printf "$LINE" | cut -f2 )
    stop=$( printf "$LINE" | cut -f3 )
    region=$chr:$start-$stop
    echo "Regions:" $region 
    echo "start:" $start
    echo "stop:" $stop
    echo "line:" $LINE

    mkdir -p results/${region/-/_}

    samtools view -F 4 $inBAM $region  | while read L; do (samtools view -H $inBAM && echo "$L" ) | samtools view -bS - | samtools mpileup -f $refGenome -r $region - > results/${region/-/_}/${FILE%.bam}Mpileup.txt; done
done < $BED

unfortunately this fails. I need to have the mpileup files inside a different region directory so that I can distinguish them. Later on and i will not be left over with 1000 files if I have 200 regions in my bed file. Hope this is also possible thanks for you help already!

this is the error I got:

[bam_header_read] EOF marker is absent. The input is probably truncated.
[samopen] SAM header is present: 93 sequences.
[bam_index_load] fail to load BAM index.
[mpileup] fail to load index for -