fix insufficient MaxProbPropagationDistance

[TomofumiSaka]

Dear GATK Team,

I may have found code with unintended behavior in the HaplotypeCaller.
The field MaxProbPropagationDistance of the ActivityProfile class can be set in the --max-prob-propagation-distance option and defaults to 50.
The field MAX_FILTER_SIZE of the BandPassActivityProfile class is 50.
I guess that the default value for --max-prob-propagation-distance is set as a value well above the MAX_FILTER_SIZE of the BandPassActivityProfile.

In the findEndOfRegion method of the ActivityProfile class, it is checked that sufficient length of activity profile is already calculated and added. The MaxProbPropagationDistance value is should be plus one due to the center value of gaussian filter ( please see the processState method of the BandPassActivityProfile).

You can confirm this with uploaded bam in https://pezycomputing-my.sharepoint.com/:f:/g/personal/sakai_pezy_co_jp/EkjbywcfooxJgtlT9z_vOVUBTZAfHTSvH_s1Bl1dzIH3tw?e=Y8UlB4 and following command with --max-prob-propagation-distance 51 option or this pull request change. Uploaded bam is made from open data HG001/NA12878 sample.
$java -jar $gatk HaplotypeCaller --reference Homo_sapiens_assembly38.fasta --input CNR0028194.gatk_best_practice.GRCh38.chr10_41899453-41901453.bam --output out.vcf --pcr-indel-model NONE -L chr10:41899453-41901453

This fix has very small effect to variant call. In WGS, this change affects only 10 or fewer variants.
Changing --max-prob-propagation-distance default value to 51 from 50 is also OK.