I read the wikipedia page about sequence alignment and sequence assembly but I have not been able to find any difference between the two. What is the difference between sequence alignment and sequence assembly? If there is no different why are the terminologies different?


Sequence alignment

It is done for checking sequence similarity between two or more different sequences. This will give information about how two sequences are different, what is their evolutionary relationship, which residues are conserved etc. Take a look at following sequence alignment between different sequences. (Image courtesy: Wikimedia Commons)

enter image description here

You can see conversation of different amino acid residues in variety of organisms like human, mouse, rat etc.

Sequence Assembly

It is done for creating long consensus sequence from short fragments of same sequence. Generally you would do sequence assembly after sequencing some genome or large DNA fragment. Take a look at following representative image of sequence assembly (Image courtesy: Wikimedia Commons)

enter image description here

Here black sequence is your original sequence which you wanted to sequence. However because of technical difficulties you can't sequence all at once. Hence you fragment and sequence different fragments 9which are shown in red, blue and green). Looking at overlapping region between different sequenced fragments you calculate consensus (shown in Cyan). Here you are ideally sequencing only one large sequence and assembling it.

  • $\begingroup$ but its the same thing right? I mean both could have been given the same name... $\endgroup$ – girl101 Nov 9 '15 at 8:19
  • 3
    $\begingroup$ Well, yes. Sequence assembly is kind of multiple sequence alignment ! Naming is for avoiding confusion I guess. $\endgroup$ – Dexter Nov 9 '15 at 8:21
  • 5
    $\begingroup$ @Rishika As Dexter's answer shows quite clearly, it's not the same. It would be inaccurate to say that you are doing Sequence Assembly, when you are in fact comparing genomes (doing sequence alignment) from different species (e.g. as part of a phylogenetic analysis). To say that you are doing Sequence alignment when you are doing Sequence assembly could be correct (for some steps), but slighlty misleading and not very specific. $\endgroup$ – fileunderwater Nov 9 '15 at 9:42
  • 1
    $\begingroup$ @fileunderwater is correct. I would say sequence assembly is special case of sequence alignment. $\endgroup$ – Dexter Nov 9 '15 at 9:54
  • 1
    $\begingroup$ @Dexter Again, it is not. Sequence alignment is a part of sequence assembly algorithms, but that does not make assembly a special case of alignment. $\endgroup$ – skymningen Nov 9 '15 at 12:37

The two tasks are quite different.

Aligning means comparing sequences to find differences between them.
Assembling means that you take a bunch of short sequences and try to stitch them back together to reconstitute the original DNA sequence.

Screen capture from DNA Baser Assembler of an 'assembly to reference'. The reference is in purple color. The forward and reverse sequences are red and green. The contig is blue.
Fig1: Screen capture from DNA Baser Assembler of an 'assembly to reference'. The reference is in purple color. The forward and reverse sequences are red and green. The contig is blue.

Note about the input of the two processes:

The sequences that enter into an alignment have (supposedly) already been cleaned up. Every base is correct/accurate.

The sequences that enter into an assembly are 'dirty' and inaccurate, because it is highly likely that the sequencing machine will generate 'bad' bases at the ends of the sequence. The user will have to manually inspect and clean up the sequences (even though today there are smart assembly programs that can fully automate the process reducing the time from 10 minutes per contig to basically nothing (under 1 second)). It is preferably that multiple overlapping sequences are used for redundancy. This way 'bad' bases are discovered and eliminated.

Once the assembly job is done, the clean contig can be used for further operations (such as alignment).

raw sequences
enter image description here

after reverse complement
enter image description here

final contig
enter image description here


Not necessarily. A good analogy for differentiating the two sequencing strategies (not to be confused with sequencing technologies i.e. illumina, smrt, oxford nanopore etc...) would be a puzzle. In the case of sequence alignment, you would have a bunch of puzzle pieces i.e. your reads and a picture of the finished/completed puzzle i.e. your reference genome to guide you in putting together the puzzle. In sequence assembly, you only have the puzzle pieces and no picture, no reference genome. This makes assembly a much harder task.

In the first case, you would compare the N reads against only 1 reference genome. In the second case you would compare N reads against N "quasi" reference genomes.

  • $\begingroup$ Both comparisons are nice, but only half of the truth. I can do alignment of two puzzle pieces (pairwise alignment) without a reference. And I can do reference guided assembly (what you describe is de novo assembly), by using alignment to a closely related reference as a guide for assembly. ( so assembly with a picture that does not show the current puzzle, but a similar one). $\endgroup$ – skymningen Nov 9 '15 at 12:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.