Abyss Output File Formats

Question

[Rob Syme]

There are a series of output files produced by ABYSS that describe how the contigs and scaffolds were constructed. The types of files seems to differ from version to version, so I thought it would be great to have an exhaustive list of the file format definitions.

Files that people I've seen include:

• .adj
• .dist
• .path
• .pair
• overlap.fa

What data do these files contain, and what format do they take?

[Eric Normandeau]

Hi @Rob, please consider editing into a question format. As is it, your post is really a comment and it is hard to see how we can either contribute or help you here. Cheers

Answer 1

[Rob Syme]

I'll start, but it won't be a complete answer. These descriptions are based on an answer from Shaun on the ABySS mailing list in Oct09

dist

DistanceEst --help Estimate distances between contigs using paired-end alignments.

21 : | 420877,-30,47,5.5,0
39 : 125053,-30,17,9.2,1 | 10061,-30,25,7.6,1 132861,47,16,9.5,0 219341,45,15,9.8,0 294681,44,16,9.5,0 300788,43,17,9.2,0 419176,44,16,9.5,0 426650,46,15,9.8,0
110 : 44048,-30,30,6.9,0 |
170 : 397327,9,20,8.5,0 410536,23,43,5.8,1 50658,-5,11,11.4,0 | 161076,51,11,11.4,1 283551,-27,14,10.1,0 324935,-13,25,7.6,1 4522,81,10,12.0,0
171 : 298574,39,10,12.0,1 313291,-7,20,8.5,1 | 134144,22,27,7.3,1
173 : 119549,-15,59,4.9,0 129210,-30,26,7.4,1 | 181775,-11,37,6.2,1 200470,-30,13,10.5,1

• id0 : id1,distance,number-of-pairs,error,strand | id2,...
• id1 comes after id0
• id2 comes before id0

I'm not sure what it means when there are multiple sets of space-separated groups (e.g. the 170: entry)

adj

AdjList --help Find all contigs that overlap by exactly k-1 bases.

0 [ 396267,1 ] [ 55802,0 ]
1 [ 35373,0 102691,0 ] [ 249534,0 ]
2 [ 277705,0 ] [ 350054,1 ]
3 [ 132932,0 ] [ 271245,0 182962,0 ]

• id0 length [ id1,strand ... ] [ id2,strand ... ]
• id1 comes after id0
• id2 comes before id0

overlap.fa

Overlap --help Find overlaps between blunt contigs that have negative distance estimates. Output the small contigs that fill in the gaps.

>47 25 0 33+ 38+ -15

• id0 length coverage (always 0) id1 id2 overlap
• id1 comes before id0
• id2 comes after id0
• id1 and id2 overlap by the specified number of bases (negative distance indicates overlap)

path

SimpleGraph --help Find paths through contigs using distance estimates.

@ 19,1 -> 416835,0
@ 37,0 -> 123886,1
@ 38,0 -> 401731,1 4851,0
@ 109,0 -> 43659,0
@ 168,0 -> 50189,0 69838,0 202180,1 117665,1 393524,0 406599,1

• id0,direction -> id1,strand ...
• direction = 0: id1 comes after id0
• direction = 1: id1 comes before id0

Answer 2

[Madalina]

how did obtain those files? because when I run ABySS, I obtain only the contig file?

[Rob Syme]

This question was based on an older version of ABySS. The newer versions have different outputs. I'll probably end up deleting this entry to save future confusion.

Answer 3

[Kaushik]

The above explained files are obtained when we provide paired end files

Answer 4

[virajbdeshpande]

Here are my guesses after I tried to look at the source code:

.adj files overlapindex length coverage(total count of mapping base-pairs) ; contigs connected to head [-d overlap size] ; contigs connected to tail

I do not know what is the convention for head and tail or +/- end.

Answer 5

[Nanderson-coder]

Output files of ABySS

1.${name}-contigs.fa The final contigs in FASTA format

2.${name}-bubbles.fa The equal-length variant sequences (FASTA)

1. ${name}-indel.fa The different-length variant sequences (FASTA)

4.${name}-contigs.dot The contig overlap graph in Graphviz format

Intermediate output files of ABySS

1. .adj: contig overlap graph in ABySS adj format

6 .dist: estimates of the distance between contigs in ABySS dist format

7 .path: lists of contigs to be merged

8 .hist: fragment-size histogram of a library

9 coverage.hist: k-mer coverage histogram