Does demultiplex step trim reads?

[Sanjeev Sariya]

Hi QIIME developers,

I'm working on Illumina paired end 300bp reads. I join the reads with ea-utils.

Post demultiplexing stats for read length looks like:

Median length: 438

Min length: 318

Max length: 538

Mean length 438

Command used:

R1=seqs.fq

R2=R2R3.paired.fq

split_libraries_fastq.py -i $R1 -o ./demultiplex -b $R2 -m ./mapping_qiime.txt --max_barcode_errors 1 --sequence_max_n 0 --phred_quality_threshold 20 --max_bad_run_length 300 --min_per_read_length_fraction 0.75 --barcode_type 16

I'm providing .75 as the percent for min. per read length for consecutive high quality case calls. 

Questions:

1- At what point QIIME trimmed read to 318?

2- By what metric is QIIME trimming reads?

Kindly help me to understand this.

Thanks in advance.

 

[Jose Antonio Navas Molina]

Hi Sanjeev,

Qiime uses a sliding window to check the quality of the reads. Multiple steps are performed during the quality control of the sequences. Since I don't wanna do a bad summary of the algorithm, I think the safest is to point you to the actual article that describes the algorithm: http://www.nature.com/nmeth/journal/v10/n1/full/nmeth.2276.html

Figure 1 is a good summary on how the quality control works in Qiime.

Cheers,