Nanopolish and pilon are calling wrong bases

[iltisanni]

Hi,

we sequenced and assembled a pichia genome with canu and afterwards polished it with nanopolish and with pilon+illumina reads.

We noticed that the nanopolish Assembly grows against the canu assembly and that it grows again after using pilon and illumina reads for polishing.

We then viewed our Assembly together with mapped illumina and nanopore reads in IGV and noticed that canu called some bases which nanopolish and pilon removed while polishing but those bases are correct since we can see them in the reads in IGV.

Why do nanopolish and pilon remove those bases although they are visible in both, illumina and nanopore reads...

commands we used for nanopolish and pilon:

Nanopolish:
nanopolish index -d PATH/TO/FAST5FILES -s PATH/TO/SEQUENCINGSUMMARY.TXT /PATH/TO/COMBINEDFASTQ

~/minimap2-2.10_x64-linux/minimap2 -d draft.mmi trimmed_contig.fasta
~/minimap2-2.10_x64-linux/minimap2 -ax map-ont -t 16 draft.mmi /PATH/TO/COMBINED.FASTQ | samtools sort -@ 16 -o reads_vs_trimmed_contig.sorted.bam
samtools index reads_vs_trimmed_contig.sorted.bam

python3 /usr/lib/nanopolish/nanopolish_makerange.py /PATH/TO/TRIMMED_CONTIG.FASTA | parallel --results nanopolish.results -P 3 nanopolish variants --consensus polished.{1}.fa -w {1} -r /PATH/TO/COMBINED.FASTQ -b /PATH/TO/READS_VS_TRIMMED_CONTIG.SORTED.BAM -g /PATH/TO/TRIMMED_CONTIG.FASTA -t 4 --min-candidate-frequency 0.1 --faster --fix-homopolymers

python3 /usr/lib/nanopolish/nanopolish_merge.py polished*.fa > polished_GENOME.fa

pilon:
java -jar pilon-1.22.jar \
--genome test.fasta \
--bam XXX.sorted.bam \
--fix "bases" \
--threads 16 \
--output test \
--outdir OUT