Nephele Pipeline Notes
DADA2 16S
- Details page
- Logfile has all the important messages.
- Quality profiles:
- green is the mean quality score, orange are the quartiles
- greyscale heat map
- Next step is filtering and trimming.
- Trimmed 10 bp from left of read, filtered reads with maxEE \(<= 5\) (maximum “expected errors” where \(EE = \sum(10^{(-Q/10)})\))
- In the logfile, look for
reads.in and reads.out table to see how many reads survived.
- Learns the error rates - the default parameters of DADA2 were optimized for Illumina data. If you have other data types, consult the DADA2 package help. In Nephele, we also have the option for Ion Torrent data where we use these parameters.
- Dereplicate and run dada to denoise the reads. Then merge the pairs. In Nephele, for efficiency, we follow the Big Data workflow that processes each sample separately.
- Filter sequences below 75bp (shorter sequences sometimes cause the pipeline to fail in the chimeral removal step). Remove chimeric sequences using “consensus” method which identifies potential “bimeras” (these are chimeras that are made up of 2 sequences that don’t belong together) in samples individually, and then uses a consensus “vote” over all samples to confirm. How many reads survive?
- Produce final ASVs in seq.fasta.
- Taxonomic assignment using rdp - k-mer based method with SILVA 132. Species are assigned by exact match only.
- Simple diversity analyses in graphs folder. Build tree with QIIME2 align-to-tree-mafft-fasttree. For more, see what you can do in MicrobiomeDB, and for more stats, see Nephele’s Downstream Analysis pipeline.
WGSA v2
- Details page
- Main logfile is less helpful here.
- TEDreads folder has fastp log files and decontamination (host read removal). How to read a Kraken2 report.
- TAXprofiles has taxonomic profiling of raw reads
- merged_tables
merged_Counts+Lineage.txt is the table used for the biom table submitted to MicrobiomeDB- diversity plots
- asmbMetaSpades has assemblies and individual sample gene annotations
- spades.log is easier to read than main log file
- AMRFinder output - SRR060021_asmb (stool sample) has some AMR genes
- PWYprofiles has functional annotation of (metabolic) pathways
