Skip to content

Latest commit

 

History

History
323 lines (173 loc) · 8.17 KB

reverse_lookup_vg_en.md

File metadata and controls

323 lines (173 loc) · 8.17 KB
Raw

vg command dictionary

The objectives

There are a variety of informative tutorials for vg such as (Main Wiki and Workshop in Portuguese). However when we do not know right vg subcommand and options for what we want to do, it is not easy to find them from these tutorials. To address this, we have organized command line examples of vg in temrs of what we want to do.

The version is v1.9.0 "Miglionico". The static binary and Docker image are available from here.

Constricuting a graph

Constructing a graph from a sequence

Construcint a graph from a reference genome sequence and valiant call information

vg construct -r ref.fa -v valiant.vcf > graph.vg

Constructing a graph from multiple sequence alignments

vg msga -f multi.fa > graph.vg

Format conversion

Converting vg format into a human readable format

vg view graph.vg > graph.gfa # Convert vg to GFA
vg view -j graph.vg > convert graph.json # vg to JSON
  • By converting into JSON format, it can be visualized the genome graph by using a genome graph browser such as MoMIG.   - Example:MoMIG     - Note: path information is mandately.

Converting GFA format to vg format

vg view -Fv hoge.gfa > graph.vg
  • [A bug in GFA parser with overlap with v1.9.0 was fixed] (vgteam/vg#1765)
  • When using assembly graph for downstream analysis, this   - If you use minia as an assembler, you can use here

Converting an assembly graph by SPAdes into vg

grep -v ^P assembly_graph_with_scaffolds.gfa | vg view -Fv - | vg mod -X 1000 - > graph.vg
  • Above example is confirmed by the SPAdes v3.11.1

Visualizing vg format with GraphML

vg view -d graph.vg | dot -Tpng -o vis.png # Converting vg format into dot format
vg view -dnp graph.vg | dot -Tpng -o vis.png # Highliting each path on the vg graph

Converting vg format into xg format

vg index -x index.xg graph.vg

Converting GAM format into JSON format

vg view -a mapped.gam > mapped.json

Using graphs

Shwoing statistics of a graph

Showing the total number of bases in a graph

vg stats -l graph.vg

Showing the number of nodes and edges in a graph

vg stats -z grpah.vg

Showing the number of paths in a graph

vg view graph.vg | grep ^P | wc -l

Retrieving the coodinate of a user-specified node on a user-specified path

vg find -n 10 -P chr1 -x index.xg # Where is the coodinate of the node ID 10 on the path chr 1.

Edit a graph

Dividing each node into the length of N bases or less

vg mod - X 1000 graph.vg > graph.1000.vg # Divinding each node into the 1000 bases or less.

Merging multiple nodes without path branch into one node

vg mod -u graph.vg > merged.graph.vg

Reassigning the node IDs

vg mod -c graph.vg > fixed.graph.vg

Extracting a graph consisting of nodes within N steps from a user-specified node

vg find -n 5 -c 10 -x index.xg> node5.dis10.vg # Extract the graph from the node whose ID is 5 to the node whose distance is 10

Extracting a graph consisting of nodes whose distances of bases from a user-specified node are less than N

vg find -n 5 -c 10 -L -x index.xg > node5.dis10.vg # Extracting the graph consisting of nodes whose distances of bases from the node 5 are less than 10(bp)

Extracting a graph consisting of nodes whose number is less than or equal to N from the specified path, e

vg find -n 5 -c 10 -p chr 1:50000-55520 -x index.xg > chr1:50000-55520.vg # chr1:50000-55520 and the nodes that are away from it by 10 Extract graph of

Merging multiple vg format files into one

vg ids -j 1.vg 2.vg # Aligning node IDs of 1.vg and 2.vg
cat 1.vg 2.vg > merged.vg

Extending the reference graph by adding the mutation information of the query sequence

vg augment -a direct grpah.vg aln.gam > aug.vg
  • From v1.10.0 onwards, Is the default of the option -a direct instead of pileup? → Reference
  • Unlike vg mod -i, it does not put path information. For the difference between these two, please refer here

Mapping

Creating a GCSA index

vg index -g index.gcsa -k 16 -b . graph.vg # Option -b specifies the directory where the temporary file is to be placed

# When memory consumption is too large,
vg prune graph.vg > prune.vg # Firstly simplifying the graph
vg index -g index.gcsa -k 16 -b . prune.vg # Then the foregoing commnad can be executed with less memory
rm prune.vg

Mapping paired-end reads

#It is assumed that xg and gcsa files exist
vg map -x index.xg -g index.gcsa -t 1 -f 1 fq -f 2.fq > mapped.gam

Calculating the read coverage of each base

vg pack -x index.xg -g mapped.gam -d > coverage.tsv

# If you want something like pileup, set the -e flag
vg pack -x index.xg -g mapped.gam -d -e > coverage.edit.tsv

Excluding unmapped reads

vg view -a mapped.gam | jq - cr 'select (.score > 0)' | vg view - aJG - > filtered.gam

Filtering mapping results by the percent identity (sequence similarity)

vg view - a mapped.gam | jq - cr 'select (.identity> = 0.95)' | vg view - aJG - > filtered.id95.gam

Showing statistis of mapping results

vg stats -a mapped.gam graph.vg
  • It is not used for calculation, but position argument is necessary

Among mapping results, projects corresponding to linear paths are projected onto a bam/sam file

vg surject -x index.xg -t 1 -b mapped.gam > mapped.bam

#You can extract only the mappings for the path specified with -p option
vg surject -x index.xg -t 1 -s -p chr1 mapped.gam > mapped.sam

Projection of bam/sam mapping results for reference to gam on genome graph with same path.

vg inject -x index.xg -t 1 mapped.sam > mapped.gam

Adding gene annotations into a graph as a path

To put the gene annotation as a path on the vg graph, first you need to convert the gene annotation into an alignment for the genome graph, then create a path and merge it into the vg graph.

Converting annotations in the BED or GFF format into the alignment file

vg annotate -b input.bed -x index.xg > annotation.gam
vg annotate -g input.gff -x index.xg > annotation.gam

Adding an alignment file to a graph as vg path

vg mod -P --include-aln annotation.gam graph.vg > mod.vg

# If you want to split the node at the breaks of the annotation, remove -P
# If you want to align node splittings with annotations, remove the -P option
vg mod --include-aln annotation.gam graph.vg > mod.vg

WIP: Extracting subgraph related information from a graph

Please be aware that there are uncertain points at present

Showing a list of nodes which are forming snarls.

vg snarls -m 1000 -r list.st graph.vg > snarls.pb
vg view -E list.st | jq '.visit[1: -1][].node_id | select (.! = null) | tonumber' | sort -n | uniq > node_list_in_ultra_bubble.txt

# Showing nodes in core regions (i.e. hub structures in a graph)
vg view graph.vg | grep ^S | cut -f 2 | grep -vwf node_list_in_ultra_bubble.txt > node_list_of_core_region.txt
  • A Snarl is a generalization of the superbubble which is a subgraph of a genome graph. For the definition of terms, please refer to Paten et al.
  • Note: there is an inconsistency (as of Aug 27, 2018) that the -m option of vm snarls only compute traversals for snarls with <= N nodes in the help message, but < according to the SourceCode

TODO

  • Story of valiant call