Basic question about reads pre-processing before alignment

[Apexy]

I agree with @GenoMax
Trimming can be potentially detrimental http://goo.gl/CG6eTl

[GenoMax]

I use AdapterRemoval 2.1.0 https://github.com/mikkelschubert/adapterremoval

Do you know if this program maintains the order of PE reads in the trimmed files (i.e. there are no orphan reads)?

[Mathgon]

I use AdapterRemoval 2.1.0 https://github.com/mikkelschubert/adapterremoval

[GenoMax]

FastQC knows about standard illumina adapters. If you are using some other adapters you can provide them to FastQC in a file with -a option (check fastqc -h for more info).

You may want to use a newer trimming program (e.g. bbduk from BBMap, trimmomatic etc) which is paired-end data aware. This way you would have less problems down the road. Options are easier to understand as well.

I would not worry about Q-trimming data unless you have some really bad quality data (< Q10) on ends and/or are planning to do de novo assemblies.

[610617109]

Hi, guys,

I found I got confused with some basic concepts in reads pre-processing. I try to google them but still didn't get a clear answer.
First is about adapters. Are the adapters used by the same platform the same? Fastqc can report the adapters content, how does it know which adapter the data use? Because I ran fastqc, and it told me that there was no adapters remaining. But in the data description, the data contains adapter sequence fragment CGACCAGCAT, and I found there truly are this kind of fragments in the reads.
When remove low quality score bases, there is a function in the fastx toolkit called fastq_quality_trimmer. It seems it just removes the low score base from 3'end until the score is higher than a threshold for each read. But sometimes the bases scores at 3'end are not good for all reads, we may trim that base for all reads. So how to decide which method to use?
And a question about fastx_clipper, it doesn't describe what it does, does it find the adapter we input and remove it and all the base after the adapter?
Thanks for your time.

Yue