MethylOmniData supplies coefficient tables, model files and reference matrices for MethylOmni. Version 0.4.0 adds the GlobalClock160 and GlobalClock12 coefficients, fixed training medians and model descriptions.
install.packages("remotes")
remotes::install_github("hyu020/MethylOmniData", ref = "v0.4.0")
remotes::install_github("hyu020/MethylOmni", ref = "v0.4.0")Install the data package before calculating clocks. Omitting ref installs the current default-branch version.
| Model | CpGs | Intercept | Training age range, years | Training samples |
|---|---|---|---|---|
| GlobalClock160 | 160 | 23.0109946392448 | 18–101 | 5,236 |
| GlobalClock12 | 12 | -7.554587457206566 | 18.09–101 | 5,212 |
Both models use beta values on the original 0–1 scale and return age in years. GlobalClock12 has its own coefficients. The two models share five CpGs, giving a union of 167 probes.
Each coefficient table contains probe, coefficient and training_median. The intercept has its own row. Model descriptions include the input scale, output unit, training-age range and imputation requirements.
library(MethylOmniData)
paths <- clock_resource_path("GlobalClock160")
weights_path <- paths[basename(paths) == "GlobalClock160.csv"]
weights <- read.csv(weights_path)
head(weights)
model_path <- methylomni_resource_path("GlobalClock_models")
models <- read.csv(model_path)clock_resource_path() can return more than one named path because a clock may have both coefficients and a model description. MethylOmni::load_clock_weights() selects the coefficient table automatically.
methylomni_data_manifest() lists resource identities, file paths, source information and licences. Other accessors include complex_resource_path(), omniage_resource_path() and reference_resource_path().
manifest <- methylomni_data_manifest()
manifest[manifest$resource_id %in%
c("GlobalClock160", "GlobalClock12", "GlobalClock_models"), ]Resource integrity can be checked after installation with validate_resource_manifest(). Such checks need only be repeated when the installation or its files change.
Calculate predictions and age acceleration through MethylOmni rather than editing the coefficient files. Refer to the MethylOmni user guide for examples, interpretation and missing-data methods.
MethylOmniData retains its GPL-3 licence. Source and distribution information for individual resources are recorded in inst/extdata/resource_manifest.csv.