FastQC Kmer, polyA, polyG, polyC, polyT...

[wb1016]

Hi all,

I got the Illumina reads from a NGS company. It is a bird genome. The quality control running by fastQC seems very weird. Too many poly kmers as attached.

Also, I have run the assembly, and the result was N50=150, no assembly at all!

I wonder if the reads of the sequencing is problematic??

Thanks in advanced!

[mastal]

What do the other FastQC plots look like, the per base sequence quality, and the per base sequence content?

Have you done any adapter or quality trimming before doing the assembly?

[wb1016]

The Per base sequence content also seems problematic, and other plots are fine, especially the Sequence Quality is very good.

Have trimmed the adapters and the low-quality reads before assembly.

What do the other FastQC plots look like, the per base sequence quality, and the per base sequence content?

Have you done any adapter or quality trimming before doing the assembly?

[mastal]

That actually looks OK, as long as the %GC content is what you expect for that genome.

[wb1016]

That's true. The GC content of the bird genome is always biased.

The only thing is the poly kmers, don't know if this was a problematic sequencing and it caused the failure of the assembly..

That actually looks OK, as long as the %GC content is what you expect for that genome.

[mastal]

Can you try reference-guided assembly with a related bird genome?

[wb1016]

I invested a lot of money to build seven genome libraries (including very large ones) for the project, it is expected to be a de novo sequencing... really upset with this problem...

Can you try reference-guided assembly with a related bird genome?

[mastal]

What assembler or assemblers have you used, and what coverage do you have? For some assemblers too high coverage also leads to problems.

[GenoMax]

It is still strange that you have a lot of poly-N type reads. Have you looked to see what % those are of the total data and as individual (%A,%G etc). What did you use as "low quality" read trim cutoff?

[wb1016]

What assembler or assemblers have you used, and what coverage do you have? For some assemblers too high coverage also leads to problems.

Soapdenovo2 was the assembler

since i have no reference to align ,so the coverage remains unclear

[mastal]

You could get a rough estimate, although it could turn out wrong. Use the genome sizes of the most closely related bird species with known genomes as a guide.

You said you had made several libraries of varying sizes. Do all the data have this problem, or is this just one data set that has this problem?