understanding make.contigs


I was able to get through an analysis and look at beta diversity for a small sample set. The significant differences in the AMOVAs were as I would expect and I believe the analysis was a success. However, the first step (make.contigs) removed three of my samples (B, C, and D), and while I’m pretty sure it will not affect my results, I wanted to better understand why this happened.

I understand that mothur takes the quality scores from each fastq and determines whether the forward and reverse reads line up. I also understand the cut-offs of quality scores for differences in the forwards and reverses (to some extend). A few of the files that I lost seemed to have so many bad quality reads that mothur threw away, so I thought I understood why I lost the samples. However, I reran a subset of of my fastqs to see if they would form contigs, including two samples (A and E) that formed in the overall analysis, and the three that didn’t (B, C, and D). This time I only lost one of the samples (sample B). So somehow I gained two more samples by running the make.contigs command on the same files… I then reran the entire dataset and still lost those three samples (B, C, and D), as I did originally.

Does the make.contigs command work by relating the different samples to each other? Is there some sort of proportional relationship to quality cut-offs?

Thanks for any help!



make.contigs doesn’t actually remove any sequences. Did you perhaps not set up your files file or oligos file correctly and so those samples weren’t found? Are you sure that the sequences for those samples were in the original fastq files?