Consensus Calling BCFTOOLS PacBio gaps

[uloeber]

Okay, I now might know why bcftools fails (to many real gaps in long reads which might be sorted out then). Any other suggestions? I do not find an up to date best practice for GATK3/4 or did anyone of you ever used freebayes for issues like that?

[uloeber]

Please help me people, is the question to simple or do the 120 readers have no idea as well?

[uloeber]

Dear community,
I mapped reads from targeted sequencing (PacBio ccs) to a reference (BMA mem - local alignment because I got inverse PCR reads). Most often the complete reference is not covered. I'd like to get one consensus sequence, since the second variant will never show up (diploid organism). My aim is to get something like this:
ATTTGATTTAGG-ATTGT-----------------ATGCTTCGTAT-T
At the moment
1. Seems like gaps are filled by the reference which I like to avoid
2. looking at one locus in IGV I found that in one position the reference has an A, my reads are 172 gap, 2 C, 2 A, 1 T and mpileup is calling a T???!

Here are the commands I use:

system "bcftools mpileup -x -P Pacbio -e 5 -f $ARGV[0].fasta $ARGV[0].sort.bam|bcftools call -m -Oz -o $ARGV[0].TESTvcf.gz";
#-M output site where REF allele is N
system "tabix $ARGV[0].TESTvcf.gz";
system "cat $ARGV[0].fasta | bcftools consensus $ARGV[0].vcf.gz > $ARGV[0].cns.TESTfa"

Could anybody help me which parameters I have to change (I also tried -M)? I tried a few but nothing gave me the output I expected.

Thanks in advance!