Minimap2: How to keep unmapped reads

[sarasilvapereira]

Hi! I want to subtract the PacBio reads that map to a given genome and keep only the unmapped reads for subsequent assembly. Can this be done?

Thanks

[lh3]

Output in the SAM format (option "-a"). You can get unmapped records in SAM.

[tseemann]

@ssilva1 minimap2 -a REF R1 R2 | samtools sort | samtools view -f 4 | samtools fastq -s R0' -1 R1' -2 R2' or something similar

(you might not need to sort step!)

[sarasilvapereira]

Thanks all, output to sam and then samtools view -f 4 worked well.

[tseemann]

@ssilva1 can you close this issue now? thanks!

[rbutleriii]

I am trying to use minimap2 to filter contaminant nanopore reads with a contaminant fasta. I tried the above, plus several similar syntaxes, but I cannot get the samtools fastq command to work.
minimap2 -ax map-ont contaminant.fasta nanopore.fastq.gz | samtools sort | samtools view -f 4 | samtools fastq -1 clean.fastq.gz

[sarasilvapereira]

Hi,

I had the same problem and “fixed" it by not piping (basically just separating samtools from the minimap2 command.

On 25 Jul 2018, at 06:38, Robert Butler [e-mail] wrote:

I am trying to use minimap2 to filter contaminant nanopore reads with a contaminant fasta. I tried the above, plus several similar syntaxes, but I cannot get the samtools fastq command to work.
minimap2 -ax map-ont contaminant.fasta nanopore.fastq.gz | samtools sort | samtools view -f 4 | samtools fastq -1 clean.fastq.gz

—
You are receiving this because you modified the open/close state.
Reply to this email directly, view it on GitHub https://github.com/lh3/minimap2/issues/163#issuecomment-407639166, or mute the thread https://github.com/notifications/unsubscribe-auth/AgYmTGDWFYfd3-VriNTjNq1EMmet6hTdks5uKARFgaJpZM4T87rs.

[rbutleriii]

It looks like from some checking, the samtools fastq cannot accept a piped input. In v1.9 it lets you know no input specified. I was able to pipe it through samtools view, but then no further. However, newer versions of samtools fastq allow you to specify the -f 4 or -F 4 flags for unmapped and mapped, respectively. So I ended up with:

minimap2 -ax map-ont contaminant.fasta nanopore.fastq.gz > nano.contam.sam
samtools fastq -n -f 4 nano.contam.sam | pigz > nano_clean.fastq.gz

[lh3]

Samtools fastq can accept a piped input. Use - or /dev/stdin as the file name like:

some_tools | samtools fastq - > output.fq

I am always hesitant to skip file names because most tools don't automatically test if stdin is available.

[rbutleriii]

Excellent, thanks for the help. So then minimap2 contaminant filtering becomes:
minimap2 -ax map-ont contaminant.fasta nanopore.fastq.gz | samtools fastq -n -f 4 - > nano_clean.fastq.gz

[tayabsoomro]

Could anybody please explain what the 4 in samtools -n -f 4 means?

[PanZiwei]

Could anybody please explain what the 4 in samtools -n -f 4 means?

"4" means unmapped reads, see more details at https://www.samformat.info/sam-format-flag