DunedinPoAm38#

Index#

Instantiate model class
Define clock metadata
Download clock dependencies
Load features
Load weights into base model
Load reference values
Load preprocess and postprocess objects
Check all clock parameters
Basic test
Save torch model
Clear directory

Let’s first import some packages:

[1]:

import os
import inspect
import shutil
import json
import torch
import pandas as pd
import pyaging as pya

Instantiate model class#

[2]:

def print_entire_class(cls):
    source = inspect.getsource(cls)
    print(source)

print_entire_class(pya.models.DunedinPoAm38)

class DunedinPoAm38(LinearReferenceClock):
    pass

[3]:

model = pya.models.DunedinPoAm38()

Define clock metadata#

[4]:

model.metadata["clock_name"] = "dunedinpoam38"
model.metadata["data_type"] = "DNA methylation"  # Paper: DunedinPoAm uses blood DNA methylation.
model.metadata["species"] = "Homo sapiens"  # Paper: The model was developed in members of the human Dunedin Study.
model.metadata["year"] = 2020
model.metadata["approved_by_author"] = "⌛"
model.metadata["citation"] = "Belsky, D.W., Caspi, A., Arseneault, L. et al. Quantification of the pace of biological aging in humans through a blood test, the DunedinPoAm DNA methylation algorithm. eLife 9, e54870 (2020)."
model.metadata["doi"] = "https://doi.org/10.7554/elife.54870"
model.metadata["notes"] = "Whole-blood elastic-net estimator of the rate of biological aging, trained at age 38 against a longitudinal 18-biomarker Pace-of-Aging composite measured over ages 26–38."
model.metadata["research_only"] = None
model.metadata["tissue"] = ["whole blood"]  # Paper: DNA methylation at age 38 was measured in whole blood.
model.metadata["predicts"] = ["pace of aging"]  # Paper: DunedinPoAm is a rate measure of how fast biological aging is occurring.
model.metadata["training_target"] = ["pace of aging"]  # Paper: Elastic net was fitted to the Pace-of-Aging composite derived from 18 biomarkers at ages 26, 32 and 38.
model.metadata["unit"] = ["biological years per chronological year"]  # Paper: A value near 1 represents one year of biological aging per calendar year.
model.metadata["model_type"] = "elastic net regression"  # Paper: Elastic-net regression was used to derive the methylation algorithm.
model.metadata["platform"] = ["Illumina 450K"]  # Paper: Feature selection and model fitting used age-38 Dunedin whole-blood methylation measured on the 450K array; EPIC was used for later external application.
model.metadata["population"] = "adults"  # Paper: The target was developed in 954 members of the same-year Dunedin birth cohort.
model.metadata["journal"] = "eLife"
model.metadata["last_author"] = "Terrie E. Moffitt"
model.metadata["n_features"] = 46
model.metadata["citations"] = 666
model.metadata["citations_date"] = "2026-07-05"

Download clock dependencies#

[5]:

os.system(f"curl -sL -o coefficients.csv https://raw.githubusercontent.com/bio-learn/biolearn/180852e2bab473303cb85da627178b1695ee9d86/biolearn/data/DunedinPoAm38.csv")

[5]:

Load features#

[6]:

df = pd.read_csv('coefficients.csv')
mask = df['CpGmarker'].astype(str).str.lower().isin(['intercept', '(intercept)'])
intercept_value = float(df.loc[mask, 'CoefficientTraining'].iloc[0]) if mask.any() else 0.0
coef_df = df.loc[~mask].reset_index(drop=True)
model.features = coef_df['CpGmarker'].tolist()

Load weights into base model#

[7]:

weights = torch.tensor(coef_df['CoefficientTraining'].tolist()).unsqueeze(0).float()
intercept = torch.tensor([intercept_value]).float()

[8]:

base_model = pya.models.LinearModel(input_dim=len(model.features))

base_model.linear.weight.data = weights.float()
base_model.linear.bias.data = intercept.float()

model.base_model = base_model

Load reference values#

[9]:

model.reference_values = None

Load preprocess and postprocess objects#

[10]:

model.preprocess_name = None
model.preprocess_dependencies = None

[11]:

model.postprocess_name = None
model.postprocess_dependencies = None

Check all clock parameters#

[12]:

pya.utils.print_model_details(model)


%==================================== Model Details ====================================%
Model Attributes:

training: True
metadata: {'approved_by_author': '⌛',
 'citation': 'Belsky, Daniel W., et al. "Quantification of the pace of '
             'biological aging in humans through a blood test, the DunedinPoAm '
             'DNA methylation algorithm." eLife 9 (2020): e54870.',
 'clock_name': 'dunedinpoam38',
 'data_type': 'methylation',
 'doi': 'https://doi.org/10.7554/eLife.54870',
 'notes': None,
 'research_only': None,
 'species': 'Homo sapiens',
 'version': None,
 'year': 2020}
reference_values: None
preprocess_name: None
preprocess_dependencies: None
postprocess_name: None
postprocess_dependencies: None
features: ['cg02582848', 'cg03730474', 'cg03922834', 'cg04480708', 'cg05227215', 'cg05513157', 'cg05575921', 'cg06133392', 'cg07045089', 'cg07185119', 'cg07986378', 'cg08376310', 'cg09349128', 'cg09404119', 'cg10727171', 'cg10919522', 'cg11574055', 'cg11674508', 'cg11897887', 'cg13074055', 'cg13121699', 'cg14485633', 'cg14775114', 'cg15018359', 'cg19422687', 'cg19510038', 'cg19743820', 'cg20451986', 'cg21079030', 'cg21370522']... [Total elements: 46]
base_model_features: None

%==================================== Model Details ====================================%
Model Structure:

base_model: LinearModel(
  (linear): Linear(in_features=46, out_features=1, bias=True)
)

%==================================== Model Details ====================================%
Model Parameters and Weights:

base_model.linear.weight: [0.16865555942058563, -0.030930202454328537, 0.10762208700180054, 0.32811808586120605, 0.20570090413093567, -0.19139723479747772, -0.2568254768848419, -0.013769027777016163, 0.002585785463452339, 0.007155308965593576, -0.3933064937591553, 0.0372597873210907, -0.6891953945159912, 0.010446280241012573, -0.014004146680235863, -0.32410237193107605, 0.06089318171143532, 0.07774978131055832, 0.0017560641281306744, -0.0569249764084816, -0.05277020111680031, 0.030463986098766327, -0.1716403365135193, -0.19539450109004974, 0.2227468192577362, -0.16681736707687378, -0.10328859090805054, 0.01552193146198988, 0.008607408963143826, 0.1834435760974884]... [Tensor of shape torch.Size([1, 46])]
base_model.linear.bias: tensor([-0.0693])

%==================================== Model Details ====================================%

Basic test#

[13]:

torch.manual_seed(42)
input = torch.randn(10, len(model.features), dtype=float)
model.eval()
model.to(float)
pred = model(input)
pred

[13]:

tensor([[ 2.5557],
        [ 1.7208],
        [ 1.9728],
        [ 0.4137],
        [-1.7117],
        [-1.8462],
        [-2.8839],
        [-2.1433],
        [-0.0711],
        [-1.1731]], dtype=torch.float64, grad_fn=<AddmmBackward0>)

Save torch model#

[14]:

torch.save(model, f"../weights/{model.metadata['clock_name']}.pt")

Clear directory#

[15]:

# Function to remove a folder and all its contents
def remove_folder(path):
    try:
        shutil.rmtree(path)
        print(f"Deleted folder: {path}")
    except Exception as e:
        print(f"Error deleting folder {path}: {e}")

# Get a list of all files and folders in the current directory
all_items = os.listdir('.')

# Loop through the items
for item in all_items:
    # Check if it's a file and does not end with .ipynb
    if os.path.isfile(item) and not item.endswith('.ipynb'):
        os.remove(item)
        print(f"Deleted file: {item}")
    # Check if it's a folder
    elif os.path.isdir(item):
        remove_folder(item)

Deleted file: coefficients.csv