When applying Circle-Map to classcial WGS data, I may get some FASLE POSITIVE results...

[shaka-emperor]

Hi, recently I was trying to use Circle-Map to analyze WGS data on GEO. And I got 20 "my_unknown_circle.bed" files following "Identification of circular DNA using Circle Map Realign" tutorial. The whole process was quite smooth. Then I filtered the bed files with the following command:
awk '$5 >1 && $6>50 && $3-$2>10000 && $3-$2<5000000' ../${VAR}_my_unknown_circle.bed > ${VAR}.bed

Among all these ecNDAs, some ecDNAs extensively existed with very high "Circle score". Like this one:
chrX 52857692 55652526
summary.txt

However, their mean coverage was quite low compared with the split and discordant reads. And when I open this interval in UCSC genome Browser (GRCh38), I found that the head and tail DNA sequences were similar if ligated head-to-tail. Is there any chance that Circle-Map may misjudge such junctions in WGS data? And how can I fix such problems?

Many thanks~

[iprada]

Hi!

Circle-Map is designed to work circle enriched data. I can therefore not guarantee good results in classical WGS data.

I can still try to guide you a bit though:

• How are you executing Circle-Map. Which command?
• Do you have any reason for having circles under selection? It can work well in these cases
  best,

Inigo

[shaka-emperor]

Thanks a lot, ~ and here is my command

#!/bin/bash
VAR=$1

bwa mem -t 20 -q ../../ref/bwa_build_index/GRCh38.fa ../../WGS/39PT_done/raw_data/${VAR}_1_sickle.fastq ../../WGS/39PT_done/raw_data/${VAR}_2_sickle.fastq > ${VAR}_unknown_circle.sam
samtools sort -@ 20 -n -o ${VAR}_qname_unknown_circle.bam ${VAR}_unknown_circle.sam
samtools sort -@ 20 -o ${VAR}_sorted_unknown_circle_bam ${VAR}_unknown_circle.sam
Circle-Map ReadExtractor -i ${VAR}_qname_unknown_circle.bam -o ${VAR}_circular_read_candidates.bam
samtools sort -@ 20 -o ${VAR}_sort_circular_read_candidates.bam ${VAR}_circular_read_candidates.bam
samtools index -@ 20 ${VAR}_sort_circular_read_candidates.bam
samtools index -@ 20 ${VAR}_sorted_unknown_circle.bam
Circle-Map Realign -t 8 -i ${VAR}_sort_circular_read_candidates.bam -qbam ${VAR}_qname_unknown_circle.bam -sbam ${VAR}_sorted_unknown_circle.bam -fasta ../../ref/bwa_build_index/GRCh38.fa -o ${VAR}_my_unknown_circle.bed

To be honest, I am not sure whether there exists circles in these samples or not. So the selection criterion totally depends on your tutorial...