Common experience & wisdom

Was my sequencing run bad?

Some things to keep in mind here are:

• Were yours the only samples in the sequencing lane? If not, then you will expect that a smaller fraction of the total reads will map to your known barcodes.
• How many reads did you get? The forward and reverse .fastq files should be the same length. You can get its total length with wc -l and divide by 4 (but don't do this on the head node!); or, you can get the first 1000 entries (head -4000 your.fastq), check its size (ls -lah), and then extrapolate the number of reads in the big fastq.
  • New MISEQ runs should have about 25 million reads total. Does this include ones that go to phiX?
• What do your qualities look like? The BMC data includes a file ...fastqc_report.html that shows how the quality of the read depends on nucleotide position.
• For MiSeq, you should be mostly in the green up to about 100 nucleotides.

Sanity checks

After removing primers

• What fraction of your total reads were dropped?

After barcoding

• How many index reads mapped to known barcodes?
• Were the most poorly-matched index reads the most lowly abundant?