hasindu2008
commented
8 months ago Hi there, Can you please mention which version of Dorado you used and the parameters? I am wondering if this is due to read splitting. To investigate and identify where the problem is, can I suggest you do the following and tell me the stats.
Could you grab the list of read IDs in the can_pod5_merge.blow5 by using:
slow5tools skim --rid can_pod5_merge.blow5 > blow5.rid.listCheck the count:
wc -l blow5.rid.listThen get the list of readIDs from the Dorado BAM file as:
samtools view dorado.bam | cut -f1 > bam.rid.list
wc -l bam.rid.listCan you tell me the counts you get?
Then we can do the following
grep -F -f bam.rid.list blow5.rid.list > both_in_bam_and_blow5.list
grep -v -F -f bam.rid.list blow5.rid.list > in_blow5_not_in_bam.list
grep -v -F -f blow5.rid.list bam.rid.list > in_bam_not_inblow5.listGet the read counts using wc -l for the above lists.
Also, could you post a couple of reads from the dorado BAM file as well as a couple of reads from the barcode02.fastq?
jorisbalc
Greetings,
I'm trying to demux a pod5 with skipped reads. I merged the pod5 and basecalled it using dorado and demux'd it the bam into fastq files. I convert the merged pod5 with blue-crab into a blow5:
I extract the read ids from the barcode02.fastq:
Afterwards, I try to extract from the blow5 and get the following:
It manages to extract some reads, but stops at this particular read, wondering what could be the cause of a read getting lost. Any help is appreciated!
EDIT: it seems that every 4096 read batch, it extracts less and less reads, is there any way to avoid this? Why are the reads missing?