Hi @bluecard,
Welcome to the forum!
Thanks for providing your input commands. There are a couple of issues here that I'll point out and offer a solution to.
1st, it seems as though you only have one sample, which on its own is technically not a problem but it is possible that this one sample does not contain enough reads for DADA2 to create an appropriate error model. About 100,000 reads (which is the min default) is recommended for this step. That is not to say the process will not finish, it is that we don't know whether the error model fits the data well enough. If this sample is a subset of other samples from the same run I would recommend including more samples/reads and re-run or you can also use Deblur which has a static error model and does not depend on building its own.
The second issue, which is actually causing the error you are seeing is the fact that you are truncating your reads at the 50bp spot, and in most cases this means that your paired-end reads do not overlap for merging and thus they are discarded. Essentially you are left with an empty feature-table which then cannot be processed any further.
Have a look through the forum again with regards to how truncating parameters are picked so that there is sufficient overlap for merging.
Keep us posted and good luck!
1 Like