Strange per base sequence content from FastQC report

[windnature03]

Hi all ,

I am new to bioinformatics. I have encountered some problems with the data analysis(RNA-seq for human, Pair-end 101cycle). The per base sequence content and Kmer content look pretty strange from what i expected.
Here are my questions (please use the attached file for reference)

1.There is a sudden rise of %A around 50bp. Would it be adapter contamination? but the adapter content keeps low throughout the whole run.

2. What is the possible cause of the A/T imbalance?

3. What is the possible cause of peaks around 40-49bp from the Kmer content?

4. why the base quality drops after 50bp?

Can anyone give me some clue on these questions, it's been puzzling me for a week.

Thank you

[wdecoster]

For question 1, the increase in %A is not too bad. I think it can be explained that for some reads you start running in the polyA tail, which you might want to trim off.

[nucacidhunter]

Would you know what kit was used for library prep. Some kits adapter are different from the ones that FastQC can detect.

[windnature03]

Hi nucacidhunter and wdecoster,

TruSeq RNA Library Prep Kit v2 was used, the link below shows the overpresented adapters sequence

http://imgur.com/a/QI2JV

This is the result from bioanalyzer, the peak lies around 250-300bp. Subtracting the length of the adapter (60bp), the insert should be around 120-130bp. In my opinion, it is less likely for a adatper sequence to be read at 50 cycle .

http://imgur.com/a/LGJvb

Thank you

[nucacidhunter]

Hi all ,

I am new to bioinformatics. I have encountered some problems with the data analysis(RNA-seq for human, Pair-end 101cycle). The per base sequence content and Kmer content look pretty strange from what i expected.
Here are my questions (please use the attached file for reference)

1.There is a sudden rise of %A around 50bp. Would it be adapter contamination? but the adapter content keeps low throughout the whole run.

2. What is the possible cause of the A/T imbalance?

3. What is the possible cause of peaks around 40-49bp from the Kmer content?

4. why the base quality drops after 50bp?

1 and 2- One possible explanation is 3' bias due to input RNA low quality which has increased polyA representation.

3- Sequences TATGCCG and CGTATGC are over-represented Kmers with TATGC overlap. You might check to see if they are from a particular highly expressed gene or spike in RNA if it was used.

4- It does not seem to be library related. You can ask the sequencing centre for an explanation. They can look at other lanes in the same flow cell to see if sequencing reagent or sequencer had any issues.

[GenoMax]

4. why the base quality drops after 50bp?

It is possible that inserts in your library are smaller than what you had expected. This generally causes adapter read-through and results in Q-score drops.

Have you scanned/trimmed this data for presence of adapters? I recommend you try bbduk.sh from BBMap suite for that purpose. There are threads on SeqAnswers that will guide you on how to use bbduk. You can also use bbmerge.sh or bbmap.sh (if you have a reference genome) from the same suite to estimate your library insert size.