Because metadata are only used in step 2.5+, a mismatch between them is only detected there, which can lead to more serious problem. For exemple, forgetting a sample file that is supposed to be included in the analysis, this is probably a worst-case scenario because the denoising step must be redone to correct the error, which can be time consuming.
The idea here is to design a fast comparison system in step 1 to confirm that the manifest are matching the metadata file.
Because metadata are only used in step 2.5+, a mismatch between them is only detected there, which can lead to more serious problem. For exemple, forgetting a sample file that is supposed to be included in the analysis, this is probably a worst-case scenario because the denoising step must be redone to correct the error, which can be time consuming.
The idea here is to design a fast comparison system in step 1 to confirm that the manifest are matching the metadata file.