Asa’s R packages 📦

for bioinformatics and data analysis

Chen Peng

ZJU LSI Jianglab

2024-04-19

My R packges

 

My favorite programming language is R, I have developed some R packages:

• pcutils
• pctax
• MetaNet
• ReporterScore
• iCRISPR
• plot4fun

Hello, everyone. Today I want to share some of my R packages for bioinformatics and data analysis.

My favorite programming language is R, and I have developed some R packages. By now, I have six R packages, and 5 of them are available on CRAN.

pcutils

It contains many useful functions for statistics and visualization.

Go down 👇

The first R package I developed is pcutils, which contains many useful functions for statistics and visualization. I almost use it in every project. And I built a simple website for my packages (click the title, wait a moment), you can find the documentation and examples here. (右下角) And here is the development status of pcutils. You can see the R-CMD-check results, the download statistics, and the CRAN status. (Other packages) The other packages also linked on this website.

pctax

• It contains a basic pipeline for omics analysis.
• See the pctax Tutorial

Then comes the pctax package, which contains a basic pipeline for omics analysis, especially for metagenomics data. I have written a bookdown tutorial for this package. The pctax can do many analysis tasks, such as diversity analysis, differential abundance analysis, community assembly analysis, functional enrichment analysis, network analysis, and elemental cycling analysis.

pctax features

• Metagenomic workflow
• Rtaxonkit
• Phylogenetic tree
• Diversity analysis

Built-in steps in the micro_sbatch():

• fastp
• rm_human
• kraken2
• megahit
• prodigal
• cluster
• salmon-quant
• eggnog
• cazy
• rgi
• vfdb
• …

Taxonkit is a Practical and Efficient NCBI Taxonomy Toolkit.

You can use Taxonkit in R by pctax interface as followed:

?taxonkit_lineage

# taxonkit_list
# taxonkit_reformat
# taxonkit_name2taxid
# taxonkit_filter
# taxonkit_lca

name_or_id2df(c("562","9606"),mode = "id")

  taxid      Kingdom            Phylum                  Class
1   562  k__Bacteria p__Pseudomonadota c__Gammaproteobacteria
2  9606 k__Eukaryota       p__Chordata            c__Mammalia
                Order                Family          Genus             Species
1 o__Enterobacterales f__Enterobacteriaceae g__Escherichia s__Escherichia coli
2         o__Primates          f__Hominidae        g__Homo     s__Homo sapiens

ann_tree(taxonomy, otutab) -> tree
easy_tree(tree, add_tiplab = FALSE) -> p
some_tax <- c("p__Proteobacteria","p__Actinobacteria","p__Firmicutes",
              "p__Bacteroidetes","p__Acidobacteria")
add_strip(p, some_tax)

Unified input and output format of one type of analysis for convenient downstream

• α-diversity
  • Richness, Shannon, Simpson, Chao1, ACE, PD, Pielou, …
• β-diversity
  • Distance: Bray-Curtis, Jaccard, Unifrac, Beta MPD, Phylosor, …
  • Ordination: PCA, PCoA, NMDS, PLS-DA, t-SNE, UMAP, LDA, RDA, CCA, db-RDA, …
  • PERMANOVA: adonis, anosim, mrpp, mantel, envfit, …
• ζ-diversity

Most of my projects are metagenome-related, so pctax includes a workflow of upstream analysis processes that I commonly use, including quality control, remove-host, species classification, functional annotation, etc. Using the micro_sbatch function, we can get these common scripts.

We know the Taxonkit is a Practical and Efficient NCBI Taxonomy Toolkit. pctax also includes an interface for easy use of taxonkit in R. For example, this function can be used to query the classification information using the taxonomy id.

With a table of classification, pctax can convert it into a tree object. Here is a simple example of a taxonomy tree.

pctax also contains some general analysis methods for various types of omics. When developing these parts, I tried to maintain a unified input and output format to facilitate subsequent analysis. Such as diversity analysis, including α, β, ζ diversity.

pctax features

• Differential abundance analysis
• Community assembly
• Elemental cycling

• t.test
• wilcox.test
• DESeq2
• edgeR
• limma
• randomForest
• LEfSe
• …

• Solan NCM
• Stegen (βNTI & RCbray-based)
• ST, NST, MST (stochasticity ratio)

The pctax package includes some built-in cycling pathways, which enables users to conveniently visualize all genes or KO (KEGG Orthology) involved in these pathways.

There is also differential abundance analysis, which includes a variety of methods, such as T.EST, wilcox.test, DESeq2, edgeR, limma, randomForest, LEfSe, etc.

Community assembly analysis included Solan NCM, Stegen (βNTI & RCbray-based), ST, NST, MST (stochasticity ratio), etc.

pctax includes some built-in cycle pathways that allow users to visualize all genes or KO (KEGG Orthology) involved in these pathways.

MetaNet

• It is a comprehensive network analysis package for omics data.
• See the MetaNet Tutorial

MetaNet is a comprehensive network analysis package for omics data. I also wrote a detailed bookdown tutorial for it. And here is the overall development framework of MetaNet, including data processing, network construction, network analysis, network visualization and other functions.

MetaNet features

• Support for integrated analysis for multi-omics data

• Example code
• Output figure

library(MetaNet)
multi_net_build(list(
  Microbiome = micro,
  Metabolome = metab,
  Transcriptome = transc
)) -> multi1
# set vertex_class
multi1_with_anno <- c_net_set(multi1, micro_g, metab_g, transc_g, 
                              vertex_class = c("Phylum", "kingdom", "type"))
# set vertex_size
multi1_with_anno <- c_net_set(multi1_with_anno,
  data.frame("Abundance1" = colSums(micro)),
  data.frame("Abundance2" = colSums(metab)),
  data.frame("Abundance3" = colSums(transc)),
  vertex_size = paste0("Abundance", 1:3)
)
plot(multi1_with_anno)

MetaNet has some unique features. For example, it supports integrated analysis for multi-omics data. Here is an example code for building a multi-omics network.

MetaNet features

• Calculate correlation network quickly, accelerate lots of analysis by parallel computing

• Time & Memory
• Packages comparison

secondly, MetaNet can calculate correlation network quickly, accelerate lots of analysis by parallel computing. Time, memory consumption, and saved file size approximately increase linearly with the square of the number of features. The calculation speed of MetaNet is much faster than other R packages, especially when the number of features is large.

MetaNet features

• Offer various layout algorithms for multi-omics network and interfaces to other software (Gephi, Cytoscape, ggplot2), easy to visualize

• Basic layout
• as_polygon()
• as_circle_tree()
• Group layout
• Pie node
• ggplot2
• networkD3

Layout methods for c_net_layout()

data("otutab")
pie_value=hebing(otutab,metadata$Group)
c_net_plot(multi1_with_anno,vertex.shape=c("Microbiome"="pie"),
           pie_value = pie_value,vertex.label=NA,pie_legend = T,color_legend = F)

Plot a network with pie node

as.ggig(multi1_with_anno) -> ggig
plot(ggig)

Plot a network in ggplot2 style

netD3plot(multi1_with_anno)

Plot a network in NetworkD3 style

Layout is an important part of network visualization, a good layout will present information clearly. MetaNet offers various basic layout algorithms. The as_polygon() is interesting: it can draw a network in a polygon shape, and you can change the edge numbers of the polygon. The as_circle_tree() can draw a network in a circle tree shape.

Beside the basic layout, we provide an advanced layout method for a network with group variable (such as omics or modules). It is easy to use g_layout() to control each group position and layout of each group.

And we can plot a network with pie node, which can show the composition of each node.

If you are more familiar with ggplot2, use the function as.ggig() transfer the basic R plot to ggplot2 style.

NetworkD3 can produce interactive network plot based on JavaScript

MetaNet features

• Provide comprehensive topology indexes calculation，including ecological network stability

• Network indexes
• Node indexes
• Modules
• Keystone
• Stability

test_module_net <- module_net(module_number = 3, n_node_in_modu = 30)
plot(test_module_net, mark_module = T)

The topological role of individual node

	Zi	Zi>2.5
Pi	peripherals	module hubs
Pi>0.62	connectors	network hubs

• Robust test
• Robustness
• Vulnerability
• Cohesion
• …

MetaNet can calculate comprehensive topology indexes, including network indexes and node indexes. MetaNet can also do module analysis. And use Zi-Pi analysis to find keystone nodes in a network. The stability analysis includes robust test, robustness, vulnerability, cohesion, and so on.

ReporterScore

I developed the improved ‘Generalized Reporter Score-based Analysis (GRSA)’ method, implemented in the R package ‘ReporterScore’, along with comprehensive visualization methods and pathway databases.

This work was published in Briefings in Bioinformatics recently.

Here is the graphical abstract: ‘GRSA’ is a threshold-free method that works well with all types of biomedical features, such as ko, genes, chemical compounds, and microbial species. Importantly, the ‘GRSA’ supports multi-group and longitudinal experimental designs.

And you can find the detailed introduction in my blog and the video tutorial on Bilibili.

iCRISPR

CRISPR/Cas9 system analysis in genomics and metagenomics.

The iCRISPR package is an analysis tool for CRISPR/Cas9 system in genomics and metagenomics. In the beginning, I developed it for the CRISPR project of LiuZhen, it is not completed yet. We use the CripsrCasFinder to predict the CRISPR array in the genome, but its output is not easy to use. iCRISPR can help process the output and transform to a crispr object. The crispr object containg the information of CRISPR, Cas, Array, and Spacer. We can do statistics, structure analysis, and source-target analysis with it.

plot4fun

It provides some interesting functions for plotting.

How to bend your text? How to draw a DNA? How to handle the relationship between layers? How to make a LED display?

clock(rotate_text = T)

DNA_plot()

chunlian(words = 1)

Olympic_rings()

make_LED("一起来画图！")

The last package is plot4fun, which provides many interesting functions for plotting. For example, how to bend your words? we can make a clock with the bendable text. And we can use the DNA_plot function to draw a DNA. We can use the chunlian function to draw a Chinese couplet. How to use ggplot to draw a standard Olympic rings as the Olympic rings is nested, we need to handle the relationship between layers to display the correct graphics. And here is a LED display, I used Image processing technology converts text into heatmaps and combines a GIF.

More Resources

R tips for advanced use

Official account

I have introduced some R tips for advanced use in previous lab meeting, here is the slides. For more resources, you can visit my official account, where I share my R packages and other resources.

Thanks for your attention! 🙏