IndelRealigner produce huge bam

Hi. I am using IndelRealigner for local indel realignment. The bam used as input is 6.6GB, while the realigned bam is 22GB.

Did I miss anything there?

The pipeline I used is as below:

echo "Patient ${sample}: @create intervals for local realignment"
sudo java -Djava.io.tmpdir=${out_dir}/tmpdir \
    -Xmx${maxMem} -Xms${minMem} \
    -jar ${gatk} \
    -T RealignerTargetCreator \
    -I ${out_dir}/${input_next} \
    -o ${out_dir}/${input_next}.forRealigner.intervals \
    -R ${reference} \
    -L ${intervals} \
    --interval_padding 200 \
    -rf ${reads_filter} \
    -known ${kg_mills} \
    -known ${kg_indels} \
    -nt ${maxDataThread} \
    --allow_potentially_misencoded_quality_scores \
    2>${out_dir}/logs/${sample_prefix}_createIntervals.err


echo "Patient ${sample}: @local realignment"
sudo java -Djava.io.tmpdir=${out_dir}/tmpdir \
    -Xmx${maxMem} -Xms${minMem} \
    -jar $gatk \
    -T IndelRealigner \
    -I ${out_dir}/${input_next} \
    -o ${out_dir}/${sample_prefix}.dedup.realigned.bam \
    -R ${reference} \
    -targetIntervals ${out_dir}/${input_next}.forRealigner.intervals \
    -rf ${reads_filter} \
    -known ${kg_mills} \
    -known ${kg_indels} \
    -compress 0 \
    -LOD 0.4 \
    --consensusDeterminationModel USE_READS \
    --allow_potentially_misencoded_quality_scores \
    2> ${out_dir}/logs/${sample_prefix}_realignment.err

Thanks.

Tagged:

Answers

Sign In or Register to comment.