Question: What Is The N50 Contig Number (Not N50 Length)?
6
gravatar for Rduncan
6.7 years ago by
Rduncan60
Miami, FL
Rduncan60 wrote:

I understand the meaning of N50 contig length, but lately I've been coming across assembly statistics that report the N50 contig number. Can someone explain what is meant by N50 contig number and how it relates to assembly quality?

assembly • 7.0k views
ADD COMMENTlink written 6.7 years ago by Rduncan60
8
gravatar for Jeremy Leipzig
6.7 years ago by
Philadelphia, PA
Jeremy Leipzig18k wrote:

N50-length is the length of that last contig that puts you over the top. N50-contig number (i.e. N50) is just the rank of that contig.

91 77 70 69 62 56 45 29 16 4[?]N50 vs N50 length[?] Technically N50, as opposed to N50 length, refers to the ordinal of that last contig that pushes it over the brink - in this example 4 (since 69bp is the 4th largest contig). Unfortunately, a higher N50 implies the opposite of a longer N50 length. Some papers refer to N50 length as L50, while most have simply followed the lazy convention of dropping "length" off of "N50 length". I think it is important to include units with your N50 to minimize confusion.

http://jermdemo.blogspot.com/2008/11/calculating-n50-from-velvet-output.html

ADD COMMENTlink written 6.7 years ago by Jeremy Leipzig18k

Your very good blog post is the only source I can find for the definition of N50 number. Do you recall any other sources of the definition?

ADD REPLYlink written 6.7 years ago by Neilfws48k

i don't remember how I derived the L50/N50 distinction other than deduction. Google scholar is not helping me find many L50 references. I see these terms used in the sorghum genome paper, but my blog post predates that. I see the Assemblathon http://assemblathon.org/assemblathon-1-results has actually switched the terms such that L50 is now a rank and N50 is a length - this is cute but incorrect.

ADD REPLYlink written 6.7 years ago by Jeremy Leipzig18k

It seems like this is a slightly different, and perhaps more precise, way of saying what other people have answered: that it is the number of contigs >= N50 length.

ADD REPLYlink written 6.7 years ago by Rduncan60

right, the only boundary condition really being in the rare situation where if you have a tie at the brink, let's say these are your contigs:

91 77 70 69 69 56 45 29 16 4

I would say N50 is 4, others some might say 5

ADD REPLYlink written 6.7 years ago by Jeremy Leipzig18k

OK, I see. Thanks for your answer!

ADD REPLYlink written 6.6 years ago by Rduncan60
4
gravatar for Fabian Bull
6.7 years ago by
Fabian Bull1.3k
German
Fabian Bull1.3k wrote:

Recently there seems to be a nice change in nomenclatur:

You can speak of L50 and N50:

  • L50 is the old N50 contig length.
  • N50 is the number of contigs >= L50.

At first I was very confused about the change of the meaning of N50 but now I am used to the names. It is very helpful to distinguish the different meanings.

ADD COMMENTlink written 6.7 years ago by Fabian Bull1.3k
2
gravatar for Neilfws
6.7 years ago by
Neilfws48k
Sydney, Australia
Neilfws48k wrote:

I'm guessing that N50 contig number is the number of contigs with length >= N50 length. Presumably as assembly improves, one would like total contig number to decrease and proportion of total contig number with length > N50 length to increase.

ADD COMMENTlink written 6.7 years ago by Neilfws48k

this is my understanding of N50 contig number as well, but I'm not sure I've seen this actually defined anywhere.

ADD REPLYlink written 6.7 years ago by Casey Bergman17k

No, I could not find a definition anywhere so this was my best guess. Strange, given that many assemblies are reporting the number, as the original question stated.

ADD REPLYlink written 6.7 years ago by Neilfws48k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 2257 users visited in the last hour