Host read removal

Host read removal#

There are a few different options to perform host read removal in QIIME 2: a more generic one using the filter-reads action and a more specific one using the filter-reads-pangenome action. Below you can learn how to use both of them. In this tutorial we will use the filter-reads-pangenome action to remove human reads from the dataset.

Multiple host read removal#

For cases in which a pangenome reference does not exist, or filtering reads from multiple hosts is warented, the example commands below can be “daisy-chained” (i.e. output from one filtering step is fed into the next) to remove reads from several host reference genomes. That is, you may want to remove all murine and human host reads from your samples prior to analyes. It’s often a good idea to start read removal with the host from which you are generating microbial sequence data.

For example, let’s assume you are generating microbial sequence data from the rat host, Rattus norvegicus. You’d likely start with the removal of reads from the rat reference (pan)genome, and perhaps followed by filtering against the mouse, Mus musculus (pan)genome. This will better ensure that murine host reads are removed from the sequencing data. You can then continue with host removal the human (pan)genome as well, to ensure general read removal of mamalian reads from your data.

Removal of contaminating reads#

Removal of contaminating reads can generally be done by mapping the reads to a reference database and filtering out the reads that map to it. In QIIME 2 this can be done by using the filter-reads action from the quality-control plugin. Before filtering we need to construct the index of the reference database that will be used by Bowtie 2:

  • start with the FASTA files containing the reference sequences - we will import them into a QIIME 2 artifact:

mosh tools cache-import \
    --cache ./cache \
    --key reference_seqs \
    --type "FeatureData[Sequence]" \
    --input-path ./reference_seqs.fasta

  • build the Bowtie 2 index:

mosh quality-control bowtie2-build \
    --i-sequences ./cache:reference_seqs \
    --o-database ./cache:reference_index

  • filter out the reads that map to the reference database:

mosh quality-control filter-reads \
    --i-demultiplexed-sequences ./cache:reads_trimmed \
    --i-database ./cache:reference_index \
    --o-filtered-sequences ./cache:reads_filtered

Human host reads#

Contaminating human reads can also be filtered out using the approach shown above by providing a human reference genome. Since a single human reference genome is not enough to cover all the human genetic diversity, it is recommended to use a collection of genomes represented by the human pangenome (__CIT). We have built a new QIIME 2 action filter-reads-pangenome which allows to first fetch the human pangenome sequence, combine it with the GRCh38 reference genome, build a combined Bowtie 2 index and, finally, filter the reads against it. Next to the filtered reads, the action will also return the generated index so that it can be used in any other experiments.

mosh annotate filter-reads-pangenome \
    --i-reads ./cache:reads_trimmed \
    --o-filtered-reads ./cache:reads_filtered \
    --o-reference-index ./cache:human_reference_index