benjjneb
commented
4 years ago Am I loosing too much or is it still acceptable for soil samples? And what is roughly an acceptable number of reads in general passing filter and trim?
Anything above ~25% is "acceptable" for filtering and trimming (and even below that in extreme cases. Keeping more reads is nice, but lower quality reads are less useful than you might think in improving the resolution of the sampled community, so removing them is usually doing more good than bad, even if it is reducing the raw number of reads you are working with. Here you still have tens of thousands of good quality reads passing your filter -- just fine!
termithorbor
Dear all,
I am wondering if I might loose too many reads after filter and trim. I am using V3V4 16S primers and my expected amplicon size without primers is 427. I used cutadapt to remove my primers from my sequences and this are sample quality plots:
Forward
Reverse
I used the following parameters for filter and trimm:
filterAndTrim(fnFs, filtFs, fnRs, filtRs, truncLen=c(250,200), maxN=0, maxEE=c(2,2), truncQ=2, rm.phix=TRUE, compress=TRUE, multithread=TRUE)
The result is this:
reads.in reads.out 97304 58515 71739 44064
Am I loosing too much or is it still acceptable for soil samples? And what is roughly an acceptable number of reads in general passing filter and trim?
Note: I have also tried it with maxEE=c(3,5) and now I get more reads through, but is it still reasonable?
reads.in reads.out 97304 68665 71739 51593
And is it in general recommended to analyse different sample let's say soil and water samples at the same time with DADA2 as long as the same amplicon was sequenced or is it better to do seperate analyses?
Thank you very much in advance.