I have 199 paired sequences with 4 different runs, I already have the PairedEndreads document of these sequences, but when I run the DADA2 gives me the next error. Previously, I had run-by-run analysis files, but they were with the 2022 version of qiime2. I wanted to switch to the 2024 and analyze everything together and avoid the merged of each file.
I guess it worked for you when you ran them separately. That means that with the combined dataset, your machine is running out of RAM.
Still, denoising each run is preferable:
You can save some RAM, so your machine can handle each run separately
DADA2 developers recommend denoising each run separately (with the same settings) to avoid biases introduced by different runs (related to the error model).
