Querying Data

knitr::opts_chunk$set(echo = TRUE)

```{css, echo=FALSE} pre, code {white-space:pre !important; overflow-x:auto;}

In this section we will start exploring what is stored inside the `pagoo` object and how we can access this information. Keep in mind that this object has its own associated data and methods that can be easily queried with the `$` operator. These methods allow for the rapid subsetting, extraction and visualization of pangenome data.

First of all, we will load a pangenome using a toy dataset included in the package. This is a preloaded set of 10 *Campylobacter spp.* genomes, with metadata associated.

library(pagoo) # Load package
rds <- system.file('extdata', 'campylobacter.RDS', package = 'pagoo')
p <- load_pangenomeRDS(rds)

Summary statistics

A pangenome can be stratified in different gene subsets according to their frequency in the dataset. The core genes can be defined as those present in all or almost every genome (typically 95-100%). The remaining genes are defined as the accessory genome, that can be subdivided in cloud genes or singletons (present in one genome or in genomes that are identical) and shell genes which are those in the middle. Let's see this using pagoo:


Core level

The core level defines the minimum number of genomes (as a percentage) in which a certain gene should be present to be considered a core gene. By default, pagoo considers as core all genes present in at least 95% of organisms. The core level can be modified to be more or less stringent defining the core genome. This feature exemplifies R6's reference semantics, since modifying the core level will affect the pangenome object state resulting in different core, shell and cloud sets. Have a look:

p$core_level <- 100 # Change value
p$summary_stats     # Updated object

As you can see, changing the core level from 95% to a more stringent 100% cause in decrease in the number of core genes from 1627 to 1554, and a concomitant increase in shell genes from 413 to 486. This means that 73 genes migrated from core to shell when increasing the threshold to consider a cluster as "core" to 100%. Now this changes remain in the object for subsequent analysis, or can be reverted by setting the core level again at the original value.

p$core_level <- 95

Pangenome matrix

The pangenome matrix is one of the most useful things when analyzing pangenomes. Typically, it represents organisms in rows and clusters of orthologous in columns informing about gene abundance (considering paralogues). The pangenome matrix looks like this (printing only first 5 columns):

p$pan_matrix[, 1:5]

Gene metadata

Individual gene metadata can be accessed by using the $genes suffix. It always contains the gene name, the organism to which it belongs, the cluster to where it was assigned, and a gene identifier (gid) that is mainly used internally to organize the data. Also, it may typically include annotation data, genomic coordinates, etc, but this other metadata is optional. Gene metadata is spitted by cluster, so it consist in a List of DataFrames.


If you want to work with this data as a single DataFrame, just unlist it:

unlist(p$genes, use.names = FALSE)

pagoo also includes predefined subsets fields to list only certain pangenome category, these are queried by adding a prefix with the desired category followed by an underscore: $core_genes, $shell_genes, and $cloud_genes. These kind of subsets are better explained in the '4 - Subets' tutorial, and also apply to other pangenome data described below.

Clusters metadata

Groups of orthologues (clusters) are also stored in pagoo objects as a table with a cluster identifier per row, and optional metadata associated as additional columns.


Subsets also exists for this field: $core_clusters, $shell_clusters, and $cloud_clusters.


Although is an optional field (it exists only if user provide this data as an argument when object is created), $sequences gives access to sequence data. Sequences are stored as a List of DNAStringSet (a.k.a DNAStringSetList, Biostrings package), grouped by cluster.

p$sequences             # List all sequences grouped by cluster
p$sequences[["group0001"]]  # List first cluster

Note that sequence names are created by pasting organism names and gene names, separated by a string that by default is sep = '__' (two underscores). This are the same as the gid column in the $genes field, and are initially set when pagoo object is created. If you think your dataset contain names with this separator, then you should set this parameter to other string to avoid conflicts. $sequences field also has predefined subsets: $core_sequences, $shell_sequences, and $cloud_sequences.

Organism metadata

The $organisms field contain a table with organisms and metadata as additional columns if provided.


Try the pagoo package in your browser

Any scripts or data that you put into this service are public.

pagoo documentation built on Nov. 19, 2022, 1:07 a.m.