Does Nextera Library Preparation Involve Random Hexamer Priming?

[roliwilhelm]

I cannot confirm or deny whether random hexamers are used in the PCR step during the process. My guess is that there is, based on the description of the PCR primer mix as a "PCR Primer Cocktail" and the fact that when I FastQC my sequencing data, I find sequences enriched in random hexamers in the first 9 positions of many reads. The reason I am in doubt is due to the fact, in my naive understanding, I can't see the need for the random hexamers since the transposons could carry specific sequences matching a given primer set(s).

Still, it would be great if someone could confirm or deny this.

[SNPsaurus]

You will see a bias in the first 9 positions because the tagmentation is biased. Fragments preferentially tagment at a particular sequence, so that sequence is what you see over-represented at the start of reads. The bias isn't so dramatic, so the overall coverage is pretty even (although bias does exist).

[SNPsaurus]

I didn't see your further posts on the subject and the blog post. I'll read it over and see if something additional is at work here.

[SNPsaurus]

Your blog post images match pretty well what we see... some nucleotides being up near 50% frequency at some positions. We've done tagmentation with dozens of species and see similar patterns, despite (like you) them having a wide range of GC%. If you are concerned about assembly, it might be an issue in genomes with very few of the preferred sites since those sites will tagment strongly and not well at the rest of the genome. Not sure if that makes sense! You might also see a shift in fragment sizes resulting. A colleague of mine had problems with transcriptome assembly with tagmentation. Maybe some interaction between random priming plus biased tagmentation?

[kmcarr]

I cannot confirm or deny whether random hexamers are used in the PCR step during the process. My guess is that there is, based on the description of the PCR primer mix as a "PCR Primer Cocktail" and the fact that when I FastQC my sequencing data, I find sequences enriched in random hexamers in the first 9 positions of many reads. The reason I am in doubt is due to the fact, in my naive understanding, I can't see the need for the random hexamers since the transposons could carry specific sequences matching a given primer set(s).

Still, it would be great if someone could confirm or deny this.

I'll deny it. No, there are no random hexamers involved in Nextera library preparation. See my reply to your post in the other thread for a link to a paper which demonstrates the composition bias in tagmentation site.

[nucacidhunter]

“PCR Primer Cocktail” is referring to the fact that it includes two primer mixes and it is widely used in Ilumina library prep kits. It primes flow cell lawn oligos related motif of Illumina adapters. This cocktail is not used in Nextera XT kit but is used in Nextera kit. Both kits use Nextera Index primers which have lower concentration and Cocktail provides enough primers for amplification of Nextera DNA libraries which uses 50x DNA than XT libraries.

[bilyl]

I cannot confirm or deny whether random hexamers are used in the PCR step during the process. My guess is that there is, based on the description of the PCR primer mix as a "PCR Primer Cocktail" and the fact that when I FastQC my sequencing data, I find sequences enriched in random hexamers in the first 9 positions of many reads. The reason I am in doubt is due to the fact, in my naive understanding, I can't see the need for the random hexamers since the transposons could carry specific sequences matching a given primer set(s).

Still, it would be great if someone could confirm or deny this.

PCR Primer Cocktail is a primer set that primes to the P5/P7 sequences, which correspond to the ones that hybridize to the flow cell. The random "tagmentation" of DNA is achieved by the pseudo-random integration of a Tn5 transposome payload into target template DNA. This was commonly used to create mutagenesis libraries back in the day but Epicentre/Illumina did a nice trick whereby a non-continuous payload consequently causes the target DNA to become fragmented.

Epicentre/Illumina have patents that go over this in detail. In addition, their user guides and brochures are pretty open about the basic molecular biology. There are also plenty of papers from the 80s/90s that go into Tn5 biology.

I'm not really sure what you're getting at with the first 9bp analysis. Nextera/Tn5 was never claimed to be completely random, but rather sufficiently unbiased so that it doesn't matter given a large enough genomic diversity (eg. the human genome).

[roliwilhelm]

Thank you all very much for the thorough responses.

@kmcarr: I appreciate your help in answering my questions.