Odd PCoA from distance matrix

[jbisanz (Jordan Bisanz)]

Yes, therein is your problem. Assuming that the the sequencing run was set up identically for each of the 3 runs, you need to match the trim-left parameters across sequencing runs. Conceptually your problem is that when combining sequencing runs, it is done on perfect matching of the sequences, and if the length of the ends is variable, they will be treated as different sequences as below:

Run1 ..AGCGCGT...
Run2 .GAGCGCGT...
Run3 AGAGCGCGT...

Mehrbod has done a better job of explaining it here:

denoising on multiple runs to be combined down stream

Hi @hsapers. The ultimate goal here is that when you are comparing reads they should be of the exact same region, otherwise even a single nucleotide between them would lead them to be called a different feature. For example: In study 1: V4 region, trim length = 0 F: AACCGGTT In study 2: same V4 region, same read, trim length =1 F: ACCGGTT So even though these 2 reads are technically identical and should be called the same feature, with a single nucelotide trimming they will be called dif…

Unrelated issue which I see from your length distribution is the large number of features you have that are around ~250bp. I recall there is a bit of length variation in V3-V4 but I don't think that much. You may want to investigate how others are trimming and overlapping their reads for V3-V4 on MiSeq (I assume from the truncation lengths you are using).