Parameter choose when remove chimeras in DADA2

When I use q2 dada2 I found a weird default parameter
–p-min-fold-parent-over-abundance 1.0
is that too low? I believe it may cause some false negative.
Uchime2 use a default 16.0 but I think it is a little bit aggressive.
What should I choose?

1 Like

Hi @sixvable,
This is an excellent question that I look forward to hearing expert advise on.
In the mean time, selecting chimera settings can be tricky because chimera removal in itself is a rather complicated and unsolved problem. Uchime’s author Robert Edgar has an excellent pre-print on the topic that I think is worth reading with regards to this.
You are right that default 1 may be a bit conservative and lead to false negatives, but I would argue that in most cases this is a good thing. If you are feeling like too many reads are being removed as chimera you can certainly try the consensus method or try increasing min-fold-parent-over-abundance to a starting 8 for example. While not possible in q2-dada2, in native dada2 you can look through the removed chimeras and I like to blast those to see how they look before adjusting my chimera removal parameters. This may not be a bad addition to q2-dada2 to optionally save the removed chimeric sequences for viewing.

3 Likes