Is there any NGS tool which produce long pairwise reads?

Question

[manetsus]

AFIK, Short sequencing and short pairwise sequencing generated by next-generation-sequencing is common now-a-days and there exists several alignment tools for them.

But I want to ask about long pairwise sequencing. There exists some alignment tools for long sequences like rHAT. That means there is at least one tool which generates long reads.

Now, I am asking that is there any tool which generates Long Pairwise Reads? if yes, then is there exist any aligner or mapper for that? Give me some idea about it.

Thanks in advance.

[GenoMax]

If your reference to pairwise reads is referring to reading both strands of DNA at the same time .. then no such technology exists (as noted below Oxford nanopore/Pacbio can read both DNA strands sequentially, but not at the same time). In addition there is no need for this since DNA strands are anti-parallel complements of each other so once you sequence one strand the corresponding base on the other strand can be inferred.

[WouterDeCoster]

For oxford nanopore you can also read the other strand for consensus calling and higher quality. It was more important with older chemistry, but it's still an improvement with current chemistry.

[manetsus]

By pairwise reads, I meant k+d+k length reads where d is the gap between two elements of the pair. Each element consist of length k. In case of computation which is written separating a | sign. This problem may clear a little bit more about the pairwise reads. Thanks a lot for your time.

[GenoMax]

This is a bit of a nitpick but you should call these "paired end" reads, a more accurate representation of the scenario you described. A single fragment being sequenced from two ends.

[manetsus]

The word "paied" is unknown to me and I found nothing by googling. Is this a typo?

[GenoMax]

It was a typo. Thanks for the reminder.

[User 59]

Where d presumably is greater than the normal insert size for an Illumina NGS library?

Answer 1

[igor]

See this nice graphic for various sequencing technologies currently available:

from: https://flxlexblog.wordpress.com/2016/07/08/developments-in-high-throughput-sequencing-july-2016-edition/

None of the long read technologies are paired-end. However, with long reads, the paired-end option becomes a lot less useful.

[manetsus]

Thanks for your nice answer. But could you please tell me a little bit more about the last line of your answer? Is there anyway to understand in the graph, which are paired-end and which are not? Could you please give me any source about the information "None of the long read technologies are paired-end"? I believed your statement, but I have to convince other, thats why I need source. Again Thanks for short but informative answer.

[WouterDeCoster]

You don't really need paired end data if your read is already 10-20kb. It doesn't have an added value, and isn't possible with the long read technologies (Oxford Nanopore (MinION/PromethION) and PacBio (RS/Sequel))

[manetsus]

That means Oxford Nanopore (MinION/PromethION) and PacBio (RS/Sequel) never generates paired-end read, right? "You don't really need paired end data if your read is already 10-20kb. It doesn't have an added value" -- if the genome size is huge then?Then also doesn't it have an added value?

[WouterDeCoster]

It's physically not possible by the type of sequencing that is done. It's not like illumina. Essentially, Oxford Nanopore sequences an entire multi-kb fragment from 5' to 3' (and back again depending on chemistry) and PacBio sequences a circular template (multiple times, depending on the length).

Well, maybe "not having an added value" isn't completely true, but compared to Illumina it's of course far less relevant. The problem with short reads is that you cannot resolve repetitive element which are longer than your read length, but having 20kb reads will resolve many repetitive elements.

[igor]

Also, with Illumina, using paired-end reads means you are sequencing more of the DNA fragment. For example, if the fragment is 500bp long, doing a 2x100 run will sequence 100bp from both ends. Thus, you are sequencing 200bp of the fragment. There is a limit on the number of cycles. A paired-end run will have a higher total limit because there will be a second sequencing reaction for each fragment.

For newer long read technologies, as WouterDeCoster already mentioned, you are sequencing the entire fragment anyway.

[User 59]

I'm still confused by what you mean by long pairwise reads - have you (for instance) looked at the construction of long mate-pair libraries? Or 10X Genomics? I wonder if these might be of interest to you from the question..

Answer 2

[i.sudbery]

Depends on your definition of "long". The idea of paired end sequencing originally arose with clone based sanger sequencing when sequencing primers would be designed to read in from both ends of the insert. This was used with short inserts (such as a standard plasmid based cloning vector) to a) effectively extend the maximum read length, OR provide two readings of a sequence for error checking. With long insert libraries, such as BACs, paired-end sequencing allowed placement of sequence on scaffolds, and detection of structural rearrangements and copy number alterations.

[manetsus]

Thanks for your answer. Excuse me, but what is BAC?

[i.sudbery]

A BAC is a Bacterial Artificial Chromosome. They are used as cloning vectors to hold very large inserts - in the 100s of kbp, and even up to a Mbp. They were heavily used in clone by clone genome sequencing projects like the Human Genome Project, and were also used in genome engineering techniques before the advent of CRISPR.