propr

[OFFICIAL] [CoDA] An R package to calculate proportionality and other measures for compositional data

https://github.com/tpq/propr

Science Score: 26.0%

This score indicates how likely this project is to be science-related based on various indicators:

○
CITATION.cff file
✓
codemeta.json file
Found codemeta.json file
✓
.zenodo.json file
Found .zenodo.json file
○
DOI references
○
Academic publication links
○
Committers with academic emails
○
Institutional organization owner
○
JOSS paper metadata
○
Scientific vocabulary similarity
Low similarity (16.4%) to scientific vocabulary

Last synced: 11 months ago · JSON representation

Repository

[OFFICIAL] [CoDA] An R package to calculate proportionality and other measures for compositional data

Basic Info

Host: GitHub
Owner: tpq
Language: R
Default Branch: master
Homepage:
Size: 24.9 MB

Statistics

Stars: 71
Watchers: 2
Forks: 12
Open Issues: 28
Releases: 17

Created over 10 years ago · Last pushed about 1 year ago

Metadata Files

Readme Changelog

README.Rmd

---
output:
  md_document:
    variant: markdown_github
---



```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

## Attention!!

If you have used `propr` previously, you will notice some changes. In version 5.0.0, I have completed a major revision of the code base to simplify the maintenance of `propr` going forward, including a restructure of back-end and front-end functionality. From the user's perspective, you will notice a few changes. First, all supporting visualization functions are gone. They were poorly implemented and not backwards compatible. Instead, you can use the unified `getResults` wrapper to pull data from `propr` and `propd` objects to pipe to `ggplot2` for visualization. Second, many experimental functions have been removed, with the remaining ones all sharing the prefix `run`. The core routines called by the `propr` and `propd` functions remain unchanged.

## Introduction

The `propr` package provides an interface for 4 distinct approaches to compositional data analysis (CoDA): proportionality, differential proportionality, logratio partial correlation with basis shrinkage, and ratio analysis.

If you use this software, please cite our work. We don't get paid to make software, but your citations help us to negotiate support for software maintenance and development.

```{r}
citation("propr")
```

OK, now let's get started.

```{r, eval = FALSE}
counts <- matrix(rpois(20*50, 100), 20, 50)
group <- sample(c("A", "B"), size = 20, replace = TRUE)
devtools::install_github("tpq/propr")
library(propr)
```

## Proportionality

There are a few proportionality statistics available. Select one with the 'metric' argument.

```{r, eval = FALSE}
pr <- propr(
        counts,  # rows as samples, like it should be
        metric = "rho",  # or "phi", "phs", "vlr"
        ivar = "clr",  # or can use another gene as reference, by giving the name or index
        alpha = NA,  # use to handle zeros
        p = 100  # used for permutation tests
      ) 
```

You can determine the "signficance" of proportionality using a built-in permutation procedure. It estimates the false discovery rate (FDR) for any cutoff. This method can take a while to run, but is parallelizable.

```{r, eval = FALSE}
pr <- updateCutoffs(
        pr,
        number_of_cutoffs = 100,  # number of cutoffs to estimate FDR
        custom_cutoffs = NULL,  # or specify custom cutoffs
        tails = 'right',  # consider only the positive values ('right') or both sides ('both')
        ncores = 4  # parallelize here
      ) 
```

Choose the largest cutoff with an acceptable FDR.

## Logratio partial correlation with basis shrinkage

There are many ways to calculate partial correlations, with or without shrinkage. The recommended one for datasets with p>>n and influenced by compositional bias is "pcor.bshrink".

```{r, eval = FALSE}
pr <- propr(
        counts,  # rows as samples, like it should be
        metric = "pcor.bshrink",  # partial correlation without shrinkage "pcor" is also available
        ivar = "clr",  # "clr" is recommended
        p = 100  # used for permutation tests
      ) 
```

You can also determine the "significance" of logratio partial correlations with the built-in permutation approach.

```{r, eval = FALSE}
pr <- updateCutoffs(
        pr,
        number_of_cutoffs = 100,  # number of cutoffs to estimate FDR
        custom_cutoffs = NULL,  # or specify custom cutoffs
        tails = 'right',  # consider only the positive values ('right') or both sides ('both')
        ncores = 4  # parallelize here
      ) 
```


## Differential Proportionality

There are also a few differential proportionality statistics, but they all get calculated at once.

```{r, eval = FALSE}
pd <- propd(
        counts,
        group,  # a vector of 2 or more groups
        alpha = NA,  # whether to handle zeros
        p = 100,  # used for permutation tests
        weighted = TRUE  # whether to weight log-ratios
      )
```

You can switch between the "disjointed" and "emergent" statistics.

```{r, eval = FALSE}
setDisjointed(pd)
```

```{r, eval = FALSE}
setEmergent(pd)
```

You can again permute an FDR with the `updateCutoffs` method. Alternatively, you can calculate an exact p-value for $\theta$ based on a F-test. This is handled by the `updateF` method.

```{r, eval = FALSE}
pd <- updateF(
        pd,
        moderated = FALSE,  # moderate stats with limma-voom
        ivar = "clr"  # used for moderation
      ) 
```

## Getters

Both functions return S4 objects. This package includes several helper functions that work for both the `propr` and `propd` output.

```{r, eval = FALSE}
?getMatrix # get results as a square matrix
?getResults # get propr or propd results in long-format
?getRatios # get samples by ratios matrix
```

Use `getResults` to pipe to `ggplot2` for visualization.

We also provide accesory functions to get the significant pairs.
```{r, eval = FALSE}
?getSignificantResultsFDR
?getSignificantResultsFstat
?getAdjacencyFDR
?getAdjacencyFstat
?getCutoffFDR
?getCutoffFstat
```

Notice that for the getter functions to work properly on `propd` objects, you have to set the target `theta` value active:

```{r, eval = FALSE}
setActive(pd, "theta_d")
setActive(pd, "theta_e")
setActive(pd, "theta_mod")
```

## Ratio Methods

COMING SOON!!

Owner

Name: Thom Quinn
Login: tpq
Kind: user
Location: The Land Down Under

Website: https://tpq.github.io/
Repositories: 6
Profile: https://github.com/tpq

An MD in a love affair with data

GitHub Events

Total

Create event: 10
Issues event: 5
Release event: 7
Watch event: 8
Delete event: 4
Issue comment event: 10
Push event: 13
Pull request event: 14
Fork event: 3

Last Year

Create event: 10
Issues event: 5
Release event: 7
Watch event: 8
Delete event: 4
Issue comment event: 10
Push event: 13
Pull request event: 14
Fork event: 3

Committers

Last synced: over 2 years ago

All Time

Total Commits: 653
Total Committers: 4
Avg Commits per committer: 163.25
Development Distribution Score (DDS): 0.006

Past Year

Commits: 1
Committers: 1
Avg Commits per committer: 1.0
Development Distribution Score (DDS): 0.0

Top Committers

Name	Email	Commits
Thom Quinn	c**n@g**m	649
Ryan M. Moore	r**7@g**m	2
Suzanne Jin	s**n@c**l	1
suzannejin	s**o@g**m	1

Issues and Pull Requests

Last synced: 11 months ago

All Time

Total issues: 34
Total pull requests: 11
Average time to close issues: 21 days
Average time to close pull requests: 22 days
Total issue authors: 25
Total pull request authors: 4
Average comments per issue: 3.12
Average comments per pull request: 0.55
Merged pull requests: 8
Bot issues: 0
Bot pull requests: 0

Past Year

Issues: 4
Pull requests: 4
Average time to close issues: N/A
Average time to close pull requests: 1 minute
Issue authors: 4
Pull request authors: 1
Average comments per issue: 0.0
Average comments per pull request: 0.0
Merged pull requests: 4
Bot issues: 0
Bot pull requests: 0

View more stats

Top Authors

Issue Authors

jolespin (3)
suzannejin (3)
grayfall (3)
tpq (3)
JemmaSun (2)
mooreryan (2)
kmuench (1)
maubarsom (1)
handibles (1)
danfulop (1)
taylorreiter (1)
adriaaula (1)
SilasK (1)
HelloWorldLTY (1)
li000678 (1)

Pull Request Authors

suzannejin (20)
mooreryan (3)
Jayanth-kumar5566 (1)
slowkow (1)
grayfall (1)
SilasK (1)

Top Labels

Issue Labels

question (12) helpful (12) bug (5) enhancement (2)

Pull Request Labels

Packages

Total packages: 2
Total downloads: unknown
Total docker downloads: 48

Total dependent packages: 3
(may contain duplicates)
Total dependent repositories: 8
(may contain duplicates)
Total versions: 24

cran.r-project.org: propr

Calculating Proportionality Between Vectors of Compositional Data

Homepage: http://github.com/tpq/propr
Documentation: http://cran.r-project.org/web/packages/propr/propr.pdf
License: GPL-2
Status: removed
Latest release: 4.2.6
published over 6 years ago

Versions: 21
Dependent Packages: 2
Dependent Repositories: 7
Docker Downloads: 48

Rankings

Stargazers count: 6.8%

Forks count: 7.7%

Average: 10.1%

Dependent repos count: 11.8%

Dependent packages count: 14.0%

Last synced: almost 2 years ago

conda-forge.org: r-propr

Homepage: http://github.com/tpq/propr
License: GPL-2.0-only
Latest release: 4.2.6
published over 6 years ago

Versions: 3
Dependent Packages: 1
Dependent Repositories: 1

Rankings

Dependent repos count: 24.4%

Dependent packages count: 29.0%

Average: 35.0%

Stargazers count: 39.1%

Forks count: 47.5%

Last synced: 11 months ago

Dependencies

DESCRIPTION cran

R >= 3.2.2 depends
methods * depends
Rcpp * imports
fastcluster * imports
ggplot2 * imports
grDevices * imports
igraph * imports
stats * imports
utils * imports
ALDEx2 * suggests
Biobase * suggests
SDMTools * suggests
cccrm * suggests
compositions * suggests
corpcor * suggests
data.table * suggests
datasets * suggests
ggdendro * suggests
grid * suggests
knitr * suggests
limma * suggests
parallel * suggests
plotly * suggests
ppcor * suggests
reshape2 * suggests
rgl * suggests
rmarkdown * suggests
testthat * suggests
vegan * suggests

ecosyste.ms

Data

Tools

Indexes

Applications

Experiments

Open Source Science

propr

Science Score: 26.0%

Repository

Basic Info

Statistics

Metadata Files

README.Rmd

Owner

GitHub Events

Total

Last Year

Committers

All Time

Past Year

Top Committers

Issues and Pull Requests

All Time

Past Year

Top Authors

Issue Authors

Pull Request Authors

Top Labels

Issue Labels

Pull Request Labels

Packages

cran.r-project.org: propr

Rankings

conda-forge.org: r-propr

Rankings

Dependencies