We now apply filters to remove low-abundant clusters that are likely to be chimeras or artifacts. We check also if some phiX sequences are still present. Low-abundant clusters are difficult to estimate. Following FROGS guidelines, we choose 0.005% of overall abundance. More stringent filters, including filters based on the prevalence across samples, can be made later if needed.

sbatch -J filters -o LOGS/filters.out -e LOGS/filters.err -c 8 --export=ALL --wrap="module load devel/Miniforge/Miniforge3 && module load bioinfo/FROGS/FROGS-v5.0.2 && cluster_filters.py --input-fasta FROGS/SWARM/remove_chimera.fasta --input-biom FROGS/SWARM/remove_chimera.biom --output-fasta FROGS/SWARM/filters.fasta --nb-cpus 8 --log-file FROGS/SWARM/filters.log --output-biom FROGS/SWARM/filters.biom --html FROGS/SWARM/filters.html --excluded FROGS/SWARM/filters_excluded.tsv --contaminant /save/user/frogs/galaxy_databanks/phiX_genome/phi.fa --min-sample-presence 1 --min-abundance 0.00005 && module unload bioinfo/FROGS/FROGS-v5.0.2"

(to see all settings: cluster_filters.py --help)

Let look at the HTML file to see the result of cluster filters.
You have four panels: Filters by ASVs, Filters by samples, ASV distribution, and Sample distribution. Here, we will focus primarily on the Summary panel.

Since the ASV distribution and Samples distribution panels are common to multiple tools, a more detailed interpretation with vizualisation of these panels can be found in the Cluster Stat section .

Cluster filters typically remove a significant proportion of ASVs. However, these ASVs do not represent the majority of sequences.

• 99.6% of ASVs are removed, ~7% of sequences are lost but they mostly correspond to low-abundances clusters
• 213 clusters are kept!
• 962,265 sequences are remaining

You can see how many sequences have been removed by each filter (1 and 2) separately, or by both filters together, by clicking on Venn diagram (3). This will then display a Venn diagram similar to the one below. The abundance filter removed all these sequences. There did not appear to be any contaminants in our sequences.

Filters by samples

You can see details on filtered clusters within each sample. You can sort the columns by clicking on .
For example, if there had been contamination, this table would have enabled us to identify the contaminated samples.

We now apply filters to remove low-abundant clusters that are likely to be chimeras or artifacts. We check also if some phiX sequences are still present. Low-abundant clusters are difficult to estimate. Following FROGS guidelines, we choose 0.005% of overall abundance. We set the prevalence filter value to 0.5 for all samples.

sbatch -J filters -o LOGS/filters.out -e LOGS/filters.err -c 8 --export=ALL --wrap="module load devel/Miniforge/Miniforge3 && module load bioinfo/FROGS/FROGS-v5.0.2 && cluster_filters.py --input-fasta FROGS/ITS/remove_chimera.fasta --input-biom FROGS/ITS/remove_chimera.biom --output-fasta FROGS/ITS/filters.fasta --nb-cpus 8 --log-file FROGS/ITS/filters.log --output-biom FROGS/ITS/filters.biom --html FROGS/ITS/filters.html --excluded FROGS/ITS/filters_excluded.tsv --contaminant /save/user/frogs/galaxy_databanks/phiX_genome/phi.fa --min-abundance 0.00005 --min-replicate-presence 0.5 --replicate_file ITS_fast_replicates.tsv && module unload bioinfo/FROGS/FROGS-v5.0.2"

(to see all settings: cluster_filters.py --help)

This report allows to show the impact of our filters: