Hi everyone, I am new to using Qiime so i am having some hard time on perfectly picking my trunc,trim parameters from my dada2 denoising. I am curently working on 16s and i have 250bp each on forward and reverse reads, when i truncated at 245, the reads i got through the filte was very low and even when i passed 0 to the trunc so i use all my reads, i got same thing. However, when i used truncated using the 25th percentile i. 135, i got at least 40-50% of my reads passed the filter. I would appreciate knowing how this works and what the most important thing about the paramteer picks are.I attach some images below
1 Like
Hello and welcome to the forums! :qiime2:
Thank you for including so many pictures of your data and what you have already tried!
The read quality in that first plot of forward and reverse looks great!
The next pictures show some things that are less great.
- Looks like the read lengths are variable. Illumina reads are always the same length originally, but trimming program can make them variable... Was this data preprocessed before importing into Qiime2?
- The first dada2 stats file show few reads passing filter and few reads merging, which is bad...
- The second dada2 stats file show ~50% of reads passing filter (good!), but only 20-25% merging, which could be improved.
See if you can get the original Illumina data and try some more DADA2 settings to get both filtering and merging to be as high as possible!
Thanks so much @colinbrislawn . I have the demultiplexed data from the facility and i ran the fastp myself. I did not quite get the prompt
And for the one that shows 50% passing, i truncated at 135 for both reverse and forward which i am worried would leave a gap, as i am working with 16s and my expected amplicon size is 300bp. I was supposing my data looks good and would not have to truncate at any point. I have manipulated these parameters so much that i even removed sequences that have 25th percentile have quality score less than 30, yet no luck
Very interesting! 16S v4 is 250 bp with EMP primers. Is that the region you amplified?
135*2=270
So that's a 30 bp gap in a 300 bp region, but probably okay overlap for a 250 bp region.
Keep me posted on what you find!
Thanks again @colinbrislawn That's about right! i used the 515F-806R but i suppose the band can be +/- 50bp of 250. Judging with the band size i saw on my gel after PCR, that's why i am standing with the size 300bp. Let me know if there is anything else you ne
ed to know to help
I think you are on the right track!
I usually run dada2 a few times to see what works best. If I'm using new primers, I would run dada2 10+ times with different settings. It's a guess and check strategy!
OMG! I feel better knowing. i have been on it since two days. So you think i should go with the 135bp truncate?
Hopefully, you can get more than 50% reads to pass filter and join, but if that's the largest number from two days it may be time to move on...
I noticed too that you have 7 samples. Is this a trial run, or are those all your samples?
i have 14 samples in total, the remaining 7 are 18s primered, so i am just running them separately. And yes, in a way, these are like preliminary for method optimization before a large scale data collection.
1 Like