Root Exudate Feature Toolkit

Provides tools for molecule-oriented and reaction-centred analysis of root exudate datasets. It supports structural matching based on 'PubChem', calculation of molecular descriptors, and inference of candidate microbe-associated metabolic reactions using Kyoto Encyclopedia of Genes and Genomes ('KEGG') identifiers and Enzyme Commission ('EC') numbers. For background on these databases, see Kanehisa et al. (2023) and Kim et al. (2023) .

REFT

REFT (Root Exudate Feature Toolkit) is an R package for molecule-oriented analysis of root exudate and metabolomics annotation tables. It supports batch database searching, SMILES matching, and calculation of six molecular descriptors.

Main features

• Read Excel annotation tables
• Match PubChem records in the order Name -> Other_name(Kegg_name) -> Kegg_ID -> HMDB_ID
• Return a matching log and unmatched records
• Optionally export Excel output files when output_dir is explicitly supplied
• Calculate six molecular descriptors using rcdk:
  • MW
  • nHBAcc
  • ALogP
  • FMF
  • HybRatio
  • nAcid
• Query KEGG reactions linked to EC numbers for microbe-associated reaction annotation

Install dependencies

install.packages(c(
  "readxl", "dplyr", "purrr", "stringr", "tibble",
  "writexl", "webchem", "rcdk", "rcdklibs"
))

Install REFT

Option 1: Install from CRAN

install.packages("REFT")

Option 2: Install from a source tarball

install.packages("REFT_0.1.4.tar.gz", repos = NULL, type = "source")

Quick start

By default, reft_run() and reft_run_simple() return results in R and do not write files.

library(REFT)

res <- reft_run_simple(
  input_file = "example_root_exudate.xlsx"
)

head(res$descriptors)

To write output files, explicitly provide an output directory. In examples, use tempdir() or another user-chosen location.

res <- reft_run_simple(
  input_file = "example_root_exudate.xlsx",
  output_dir = tempdir()
)

Custom column names

library(REFT)

res <- reft_run(
  input_file = "your_data.xlsx",
  name_col = "Name",
  other_col = "Other_name(Kegg_name)",
  hmdb_col = "HMDB_ID",
  kegg_col = "Kegg_ID",
  output_dir = tempdir()
)

KEGG-microbe workflow

res <- reft_kegg_microbe_run(
  input_file = "microbe_ec.csv",
  output_dir = tempdir()
)

head(res$results)

Optional PubChem cache

If you want to cache PubChem results, explicitly choose a path. In examples or tests, use tempdir().

options(REFT.pubchem_cache_file = file.path(tempdir(), "REFT_pubchem_cache.rds"))

Output files

When output_dir is explicitly supplied, reft_run() writes:

• metabolites_6_descriptors.xlsx
• unmatched_smiles.xlsx
• pubchem_match_log.xlsx

When output_dir is explicitly supplied, reft_kegg_microbe_run() writes:

• microbe_ec_kegg_reactions.xlsx

Returned object

Both reft_run() and reft_run_simple() return a named list:

• descriptors: final result table containing SMILES and six descriptors
• unmatched: records that were not matched to SMILES
• match_log: PubChem matching log

reft_kegg_microbe_run() returns a named list:

• results: final microbe-EC-reaction table
• ec_to_reaction: EC-to-reaction mapping table
• reaction_details: reaction detail table
• compound_table: compound formula table

Notes

• KEGG_ID and HMDB_ID currently follow the original workflow and are queried through PubChem name-based searching, so the hit rate may vary among identifiers.
• rcdk requires a working Java environment.
• If PubChem or KEGG network access is unstable, some records may return NA.

Java / rcdk note

REFT can be installed without loading rcdk at package startup. Only the molecular descriptor calculation step requires rcdk/rJava.

If descriptor calculation fails on Windows, check that Java is installed and that R and Java use matching architectures.