Hi everyone!
I sequenced bacterial 16S via MiSeq platform at MrDNA and I got my raw reads (R1 and R2 - corresponding to F and R) and I ran quality check through FastQC. Results of sequence quality are very good, but I have more than 30 overrepresented sequences from which only one corresponds to adapter. I must say I don't have much experience in pre-processing Illumina data (but I managed process data in QIIME via standard pipeline) and I got nice diversity and not so much unassigned sequences, no matter if I used Greengenes or SILVA database.
My question is - how to trimm my sequences to get rid of those overrepresented sequences and do I need to do that at all?
Hi.
Maybe this overrepresented sequences are sequences of the phage phiX174 that usually are added to the Miseq-mix to check the correct running of the machine. You can try to blastn with nt database, to confirm (or not) that. And if also match with Amycolatopsis lurida it is the same, the phage phiX174 (both have the same sequence).
You also have to remove this kind of sequences during QIIME pipeline, discarding the no blast or unclasified OTUs after OTU picking step.
Hi Maja,
Usually the adapters are cut in the first step of analysis, and the companies provided the fastq files without the adapter. (Check it out) if you have that file. So, you have to trimm them or check in the folder with the output if those files were provided and use them for the analysis.
Another thing you found adapters different from the expected is that the adapters (as well the primers) can accumulate errors during the sequencing process.
If I was you I would first check if trimmed fastq are provided from the company.
If you don't have these files, you can try to remove the adapter.
I am not familier with QIIME, but can tell you how to do it in mothur.
Hope this is useful,
Chetan
Hi everyone,
Thank you for your answers. I found a file provided by the company that was joined reads with trimmed adapters and I checked quality with fastQC and that file also had many overrepresented sequences. Malte, I tried blasting also full sequences that represent that first 50 bases, and indeed, it seems that are some uncultured bacterial sequences. So, overall, if that large number of overrepresented sequences is just result from my dataset, do you think I should just process the fastq files just the way they are?
Hi everyone,
Just to conclude this question - I used full.fasta and full.qual files provided from MrDNA and results came out way better than with raw files!
Thanks to everyone on their help!
Hi Maja,
I am running in the same issue. What would be the full.fasta / full.qual files? Original files with adapter and before splitting? Many thanks!
Hi Raquel, sorry for the late respons, but if it still helps, all the explanations are given by the MrDNA Readme file:
1. –full.fasta, -full.qual
a. These files contain the raw sequence data information and still have primers and barcodes
Best,
Maja
- Badges
- Science topic
- Similar topics
- Correction