Interpretation of result

[sc...@ncsu.edu]

Hello, I've gained results of Abyss.

However, I am in trouble with interpreting them.

1. This table is output of stats.

	n	n:500	n:N50	min	N80	N50	N20	E-size	max	sum
contigs	128239	5402	498	500	32250	82060	183593	114516	758070	149.7e6
sdcaffolds	126745	4356	408	500	39439	101140	220518	142070	939060	149.6e6

  - What is the big difference between contigs and scaffolds?

    (Seeing the contig file and scaffold file, it does not seem that there exists a big difference between them.)

  - What are n:500, n:N50, E-size, and sum?

2. This table is output of tablet (genome viewer) of which some numbers are different from abyss.

   

Description	Value
Total number of contigs	128,239
Average contig length	1,317
Total number of reads	37,088,213 ( = 21,005,534*2 - 936,337 – 3050181 )
Average reads per contig	289
N50	70,023
N90	244
Assembly file size	555.01 MB

  

  - Why are they different?

   Abyss vs Tablet:

   (N50) 82,060 vs 70,023

   (If E-size stands for average, average contig length) 114516 vs 1317

Thank you in advance.

[Tony Raymond]

Hi,

I’ve added answers in your quoted text below.

Cheers,

Tony

Contigs are Unitigs merged using paired-end information (typically, we require there be a unique path between unitigs to make the merges), and Scaffolds are Contigs merged with either paired-end or mate-pair information (here we are only concerned with order and orientation, and fill gaps with N’s, AKA scaffolds).

    (Seeing the contig file and scaffold file, it does not seem that there exists a big difference between them.)

An N50 increase from 82kb to 101kb isn’t nothing. I would expect a greater increase if you used mate-pair data, but this is inline with using just paired-end data for scaffolding.

  - What are n:500, n:N50, E-size, and sum?

n:500 - number of contigs over 500bp long

n:N50 - number of contigs longer than the N50 length

E-size - expected size of a contig choosing a random base from the assembly.

Sum - number of ACGT characters found in contigs longer than 500bp.

2. This table is output of tablet (genome viewer) of which some numbers are different from abyss.

   

Description	Value
Total number of contigs	128,239
Average contig length	1,317
Total number of reads	37,088,213 ( = 21,005,534*2 - 936,337 – 3050181 )
Average reads per contig	289
N50	70,023
N90	244
Assembly file size	555.01 MB

  

  - Why are they different?

Looks like they are considering all contigs when calculating the statistics instead of all contigs >500bp long, which is what abyss-fac does by default. Using 'abyss-fac -s1 …’ disables the length filter and should give you the same results.

   Abyss vs Tablet:

   (N50) 82,060 vs 70,023

   (If E-size stands for average, average contig length) 114516 vs 1317

Thank you in advance.

--
You received this message because you are subscribed to the Google Groups "ABySS" group.
To unsubscribe from this group and stop receiving emails from it, send an email to abyss-users...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.