chimera removal

I wanted to know that how dada2 works?
more specifically in chimera removal… what criterion it uses to remove chimeric sequences?


Hi @Nisha,

That’s a rather broad question, and not really about QIIME 2, so might I suggest reading the original DADA2 paper for details on how this tool works.

From said paper:

Chimeras. isBimeraDenovo() identifies sequences that are exact bimeras (two-parent chimeras) of more abundant output sequences. Bimeras are identified by performing a Needleman-Wunsch global alignment of each sequence to all more abundant sequences and then searching for combinations of a left-parent and a right-parent that cover the child sequence without any mismatches or internal indels. Child sequences that differ by a single mismatch or indel from the chimeric model are also flagged if the left parent and right parent are both at least 4 nt away from the child sequence.

isBimeraDenovo() is intended to be used after denoising and on exactly inferred sample sequences, rather than on noisy input reads or fuzzy OTUs. It was necessary to implement isBimeraDenovo() because most current stand-alone chimera identification programs are intentionally conservative about identifying chimeras that are relatively close to other more abundant sequences because such sequences are expected to later be joined together in the same OTU (Supplementary Note 2). DADA2 does not create OTUs and does differentiate closely related sequence variants; therefore, we implemented this simple, but more sensitive, chimera detection method.

Hope this helps!


Thanks alot!!!

1 Like

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.