Read preparation
See also
OTU / denoising pipeline
The table below summarizes the read preparation steps that I recommend for an OTU / denoising pipeline. Follow the links for details.
| Step | Description | |
| Understand your reads | Investigate your data, don't just copy and follow an example script! | |
| Demultiplex | Assign reads to samples using index reads or barcodes | |
| Merge pairs | Merge paired reads to get consensus sequences and Q scores | |
| Strip primers | Primer-binding sequence should be removed before quality filtering | |
| Orient | If you have reads on both strands, orient before trimming and finding uniques | |
| Strip machine sequences | Remove machine-specific sequences e.g. TCAG for 454 | |
| Length trimming | Remove low-quality tails, make sure 3' ends align | |
| Quality filtering | Making OTUs (but not the OTU table) needs high-quality reads | |
| Pool samples | Reads for all samples should be combined | |
| Dereplication | Identify unique sequences and abundances | |
| Discard singletons | Remove low-abundance reads, which often have errors |
