QC on Illumina Fastq files

Question

[indranipoddar]

Hello,

I am interested to filter out contaminant, adapter, mitochondrial DNA in 3 separate fasta file from Illumina paired end fastq file. So my question is how to use bowtie2 to umap on multiple files to produce a single fastaq unmap file which will filter 3 of these.If the alternative is to use bowtie2 to map , one at atime on these 3 files, then how to combine 3 umap files as they may have redundant reads.

I highly appreciate any feedback or if this question is answered previously , then please share the link. Thanks, Indrani

[igor]

I am not sure what you mean by contaminant, but adapter and mitochondrial DNA should not map to the genome, so those will automatically be filtered out during alignment. Are you trying to remove or keep those?

[indranipoddar]

I want to remove them. I also want to remove rRNA, polyA etc. So I want to run bowte2 single run which will produce 1 single fastq file with unmapped data filtering adpater, mitochondrial, polyA, rRNA. Thanks for your reply.

Indrani

[GenoMax]

If this is RNAseq data then you don't need to remove anything except the adapters (those you don't need to strictly remove either but it is a good idea). See this recent discussion: Removing rRNA and tRNA sequences using GTF files

[indranipoddar]

Thanks for your reply.

Answer 1

[GenoMax]

If you want to capture mitochondrial reads away from the genome then you would want to look at BBSplit: BBSplit syntax for generating builds for the reference genome and how to call different builds.

Answer 2

[Brian Bushnell]

Mapping is not a good method for adapter removal, as the whole read will not match the reference. I suggest using a tool designed specifically for adapter-trimming, like BBDuk. It can also be used to remove other short contaminant sequences. For sequences longer than read length (such as mitochondrial DNA), mapping is most precise. Just concatenate all contaminant references into a single fasta file, map to it, and keep the unmapped reads. This step should be done using the reads remaining after adapter-trimming and other short synthetic contamiant removal.

[indranipoddar]

Thanks for your reply. Does bowtie2 can create index on a file with multiple fasta sequences like you suggested to concatenate all reference sequences.

Indrani

[GenoMax]

Yes it can and that is how you should do it.