Hi , can you explain in a very very simple way why the repeat sequences interfere with the assembly of the contigs in the sequencing? I can't figure it out. And why the pair-end sequencing helps to overcome this problem?I read that this involves the sequencing of both ends of a fragment, ..but I can't figure out how this process facilitates the detection of genomic rearrangements, repeats etc.... Thanks
Image you have two sequences: FS1-RS-FS2 and FS3-RS-FS4 where RS is a repeat sequence and FS1 through FS4 are some unique flanking sequences. When you try to assemble reads shorter than RS, they would only contain a portion of RS and a portion of a single flanking sequence. Hence, you will not be able to tell whether the actual sequences are FS1-RS-FS2 and FS3-RS-FS4 or FS1-RS-FS4 and FS3-RS-FS2. To resolve this, you need a read that would cover two flanking sequences. Only then you would be able to tell which two flanking sequences belong together.
That's what paired-end reads do. They allow you to connect two flanking sequences. However, the key word here is length. If you can get unpaired reads that are as long as the fragments used for paired-end sequencing, you wouldn't even need paired-end reads.
Thank you Roman for your answer..So the pared-end method is an alternative method to the normal sequencing or it is done along with the normal sequencing?
I don't think it is an alternative. Up until recently, to obtain a more or less decent draft assembly you needed to have some long-insert mate-pair reads for scaffolding. Even contig assembly benefits from paired-end reads as they can span longer repeats. However, these days more and more people are incorporating long reads, e.g. PacBio, into their assembly pipelines. I believe at this point they are mostly used for scaffolding. In principle, those reads can replace paired-end libraries eventually but their quality is not good enough yet.
A typical way of assembling contigs from shotgun reads is by breaking the reads into overlapping subsequences (so-called k-mers) of uniform length k. From the set of all these k-mers, the assembler then builds a so-called deBruijn graph. In the best case (everything sequenced, no repeats), you will be able to reconstruct the original sequence (i.e. the material that was fragmented into your short reads) from this graph; this sequence corresponds to the unique path through the graph. If you have repeats, however, the graph will not contain a single unique path anymore.
For instance:
ACGTCAGATG with k=2
kmers
AC, CG, GT, TC, CA, AG, GA, AT, TG
deBruijnGraph
AC->CG->GT->TC->CA->AG->GA->AT->TG
Sequence reconstructed: ACGTCAGATG
With repeats this looks as follows:
AAGTGCGTCA with k=2
kmers
AA, AG, GT,* TG, GC, CG*, GT, TC, CA; *) is repeat
The deBruijn Graph:
AA->AG->GT->(TG,TC); TG->GC->CG->GT->(TC); TC->CA;
When you reach GT, you have two options, either you follow TG or TC. Here, it is still rather straightforward, but with more repetitive content, you will run into trouble.
Paired-end reads can help out with this, when their insert size is long enough so that they bridge the repeat. In that case, you have additional information for each extracted kmer and can place it more accurately.
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology