Too many reads containing all N's in Miseq read

[windhorse8]

Dear All,
I recently used Miseq for small RNA library sequencing (single end, 35 cycle). The library was prepared using size selected (19-24 nt) small RNA sequences. When I analyzed the result using Galaxy FastQC, 50 % of my reads contain all Ns. What could be the reason ? Does that mean my sequencing result was horrible ? Also, in the reads that contain readable sequences, all of them are followed at the end with about 7-9 N bases. Could this be because my actual small RNA sequences were shorter than the cycle number and somehow, the instrument started calling for N once it started reading adapter sequence that follows the sRNA insert sequence ?
I am new with NGS. Any help will be highly appreciated.

Cheers.

[Brian Bushnell]

This indicates to me a high probability of a library without sufficient color-balancing. Can you post your fastQC graph of per-position base content, and another of per-position average Q-score? Also, I don't know a lot about small-RNA library-prep - does it involve custom amplification primers, or just size-selection?

[windhorse8]

Hi Brian,
Please find the dropbox link for FastQC report for my data.

https://www.dropbox.com/s/4i9642io1gk7ru0/KR3_S1_L001_R1_001.fastq%20FastQC%20Report.pdf?dl=0

The 3' adapter ligated sRNA samples were size-selected by gel fractionation and used for 5' adapter ligation followed by reverse transcription and PCR amplification.
Thanks again for your help.

Lodoe

[Brian Bushnell]

There is definitely substantial base-content bias, but enough variation that lack of color-balancing shouldn't be the problem. And even if you had 0bp sequences, Illumina primers themselves would still go out past cycle 31 - furthermore, when you go off the end of the molecule, you don't get Ns, but typically bases called from noise from nearby clusters.

Hmm... maybe someone else has a better idea of what went wrong, but it certainly looks like a failed run to me. I don't know if that's due to the machine or the library.

[GenoMax]

Did you spike phiX in this run? There could be several things wrong here. Low nucleotide diversity, bad libraries are all possibilities. Have you contacted illumina tech support and have them take a look at the raw data?

[nucacidhunter]

To my knowledge libraries prepared with TruSeq small RNA kit can not be sequenced on MiSeq due to lack of sequencing primers for them in primer mixes of kit.

[dariober]

50 % of my reads contain all Ns. What could be the reason ?

Hi- This is just an hypothesis. What is the origin of the fastq files? If they went through the MiSeq processing pipeline it might be that adapter sequences have been masked with N, hence the large number of Ns you see (see also
http://support.illumina.com/content/...15028392-j.pdf). Note also that you don't observe enrichment for adapter sequences further supporting the hypothesis that adapters have been removed from this file. Also, the raw fastq files I have seen from Illumina sequencers don't contain N no matter how bad the quality is, you get one of ACTG.

[dariober]

To my knowledge libraries prepared with TruSeq small RNA kit can not be sequenced on MiSeq due to lack of sequencing primers for them in primer mixes of kit.

Do you have more information about this? From this page it seems the MiSeq supports small RNA library preps.

[nucacidhunter]

Hi dariober
Once I tried to prepare a sample sheet with Illumina Experiment Manager for small RNA library, but there was not any option for small RNA under MiSeq. I do not know if new version have that option.

[windhorse8]

Hi Dariober,
The fastQ file was sent by our collaborator who did the sequencing. I don't know if they went through Miseq pipeline. However, based on what you are saying and info provided in the link you sent, it seems like adapter sequences are masked. The reads containing all N's could represent adapter dimer sequences. Also, the reads with the actual sequence trailed with 6-8 N's could be the sRNA sequence followed by adapter sequences which are again masked.
Thank you again for your help. I will contact our collaborator and ask more details about the sequencing protocol followed for my library.

[kmcarr]

Hi dariober
Once I tried to prepare a sample sheet with Illumina Experiment Manager for small RNA library, but there was not any option for small RNA under MiSeq. I do not know if new version have that option.

Don't know what version you used but IEM v1.8.2 certainly does.

MiSeq --> RNA Sequencing (Category) --> Small RNA (Application) if you want MiSeq Reporter to run the small RNA analysis.

OR

MiSeq --> Other --> FASTQ Only and then choose TruSeq Small RNA from the Sample Prep Kit drop down on the next page if you just want the sequence files.

[mgfield]

Reason for so many Ns

This same issue happened to me. What causes it is the minimum threshold setting for read length after adapter trimming. The default setting after trimming adapters is to label everything under 35 nucleotides to be Ns. So, if you used a read length of 75 nucleotides and tried to sequence miRNA, then the miRNA would be less than 35 nucleotides and show up all as Ns. This minimum threshold can be set by running bcl2fastq. However, if you are doing your analysis in basespace, you can either set a regular mRNA or miRNA library but not both. So, if you select a mRNA library, then the small miRNA reads will be masked as Ns. This happened to me when I combined ChIP-Seq data with miRNA data. There is NO way to fix this in basespace after the run. You have to download the raw bcl files and then convert bcl2fastq by changing the minimum threshold setting. I know this is way after you posted this, but if you couldn't figure out what was causing the problem... here it is.