DunedinPACE

DunedinPACE#

Index#

  1. Instantiate model class

  2. Define clock metadata

  3. Download clock dependencies

  4. Load features

  5. Load weights into base model

  6. Load reference values

  7. Load preprocess and postprocess objects

  8. Check all clock parameters

  9. Basic test

  10. Save torch model

  11. 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.DunedinPACE)

class DunedinPACE(pyagingModel):
    def __init__(self):
        super().__init__()

    def preprocess(self, x):
        """
        Apply quantile normalization on x using gold standard means.
        """
        # Ensure gold_standard_means is a 1D tensor and sorted
        sorted_gold_standard = torch.sort(torch.tensor(self.reference_values, device=x.device, dtype=x.dtype))[0]

        # Pre-compute the quantile indices
        quantile_indices = torch.linspace(0, len(sorted_gold_standard) - 1, steps=x.size(1)).long()

        # Prepare a tensor to hold normalized data
        normalized_data = torch.empty_like(x, device=x.device, dtype=x.dtype)

        for i in range(x.size(0)):
            sorted_indices = torch.argsort(x[i, :])
            normalized_data[i, sorted_indices] = sorted_gold_standard[quantile_indices]

        # Return only the subset from x that is used in the base model
        return normalized_data[:, self.preprocess_dependencies[0]]

    def postprocess(self, x):
        return x


[3]:

model = pya.models.DunedinPACE()

Define clock metadata#

[4]:

model.metadata["clock_name"] = 'dunedinpace'
model.metadata["data_type"] = 'methylation'
model.metadata["species"] = 'Homo sapiens'
model.metadata["year"] = 2022
model.metadata["approved_by_author"] = '✅'
model.metadata["citation"] = "Belsky, Daniel W., et al. \"DunedinPACE, a DNA methylation biomarker of the pace of aging.\" Elife 11 (2022): e73420."
model.metadata["doi"] = "https://doi.org/10.7554/eLife.73420"
model.metadata["research_only"] = True
model.metadata["notes"] = "This model is for research purposes only. Commercial users should contact exclusive DunedinPACE licensee TruDiagnosticTM. The automatic failure if fewer than 80% of the CpG probes are available is not implemented and left to the user's discretion."

Download clock dependencies#

Download from R package#

[5]:

%%writefile download.r

options(repos = c(CRAN = "https://cloud.r-project.org/"))
library(jsonlite)

install.packages("devtools")
devtools::install_github("danbelsky/DunedinPACE", build_vignettes = FALSE)
library(DunedinPACE)

PACE_list = list(c(
    mPACE_Models$model_names,
    mPACE_Models$gold_standard_probes,
    mPACE_Models$model_weights,
    mPACE_Models$model_intercept,
    mPACE_Models$model_means,
    mPACE_Models$model_probes,
    mPACE_Models$gold_standard_means
))

write_json(PACE_list, "DunedinPACE.json", digits = 12)

Writing download.r

[6]:

os.system("Rscript download.r")

[6]:

0

Load features#

From JSON file#

[7]:

with open('DunedinPACE.json', 'r') as f:
    PACE_list = json.load(f)[0]
model.features = PACE_list['DunedinPACE']
model.base_model_features = PACE_list['DunedinPACE.4']

Load weights into base model#

[8]:

weights = torch.tensor(PACE_list['DunedinPACE.1']).unsqueeze(0).float()
intercept = torch.tensor([PACE_list['DunedinPACE.2'][0]]).float()

Linear model#

[9]:

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#

[10]:

model.reference_values = PACE_list['DunedinPACE.5']

Load preprocess and postprocess objects#

[11]:

model.preprocess_name = 'quantile_normalization_with_gold_standard'

indices = [model.features.index(item) for item in model.base_model_features]
model.preprocess_dependencies = [indices]

[12]:

model.postprocess_name = None
model.postprocess_dependencies = None

Check all clock parameters#

[13]:

pya.utils.print_model_details(model)


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

training: True
metadata: {'approved_by_author': '✅',
 'citation': 'Belsky, Daniel W., et al. "DunedinPACE, a DNA methylation '
             'biomarker of the pace of aging." Elife 11 (2022): e73420.',
 'clock_name': 'dunedinpace',
 'data_type': 'methylation',
 'doi': 'https://doi.org/10.7554/eLife.73420',
 'notes': 'This model is for research purposes only. Commercial users should '
          'contact exclusive DunedinPACE licensee TruDiagnosticTM. The '
          'automatic failure if fewer than 80% of the CpG probes are available '
          "is not implemented and left to the user's discretion.",
 'research_only': True,
 'species': 'Homo sapiens',
 'version': None,
 'year': 2022}
reference_values: [0.8499678448731, 0.7897610893879, 0.9717597609504, 0.8480138104804, 0.2475593905826, 0.1097207575569, 0.189885225724, 0.3909313843346, 0.9168688056148, 0.2708397202163, 0.8736696325841, 0.5147248840362, 0.9619479966112, 0.2575877641187, 0.9432588276732, 0.8989191043041, 0.580889869872, 0.1242693996089, 0.8932207211244, 0.6638228904263, 0.09830894494804, 0.8640316497494, 0.08535902315207, 0.07942500721274, 0.7473024928443, 0.8737263989611, 0.7534216033511, 0.09034449086512, 0.109486844368, 0.6611683137784]... [Total elements: 20000]
preprocess_name: 'quantile_normalization_with_gold_standard'
preprocess_dependencies: [[0,
  1,
  2,
  3,
  4,
  5,
  6,
  7,
  8,
  9,
  10,
  11,
  12,
  13,
  14,
  15,
  16,
  17,
  18,
  19,
  20,
  21,
  22,
  23,
  24,
  25,
  26,
  27,
  28,
  29,
  30,
  31,
  32,
  33,
  34,
  35,
  36,
  37,
  38,
  39,
  40,
  41,
  42,
  43,
  44,
  45,
  46,
  47,
  48,
  49,
  50,
  51,
  52,
  53,
  54,
  55,
  56,
  57,
  58,
  59,
  60,
  61,
  62,
  63,
  64,
  65,
  66,
  67,
  68,
  69,
  70,
  71,
  72,
  73,
  74,
  75,
  76,
  77,
  78,
  79,
  80,
  81,
  82,
  83,
  84,
  85,
  86,
  87,
  88,
  89,
  90,
  91,
  92,
  93,
  94,
  95,
  96,
  97,
  98,
  99,
  100,
  101,
  102,
  103,
  104,
  105,
  106,
  107,
  108,
  109,
  110,
  111,
  112,
  113,
  114,
  115,
  116,
  117,
  118,
  119,
  120,
  121,
  122,
  123,
  124,
  125,
  126,
  127,
  128,
  129,
  130,
  131,
  132,
  133,
  134,
  135,
  136,
  137,
  138,
  139,
  140,
  141,
  142,
  143,
  144,
  145,
  146,
  147,
  148,
  149,
  150,
  151,
  152,
  153,
  154,
  155,
  156,
  157,
  158,
  159,
  160,
  161,
  162,
  163,
  164,
  165,
  166,
  167,
  168,
  169,
  170,
  171,
  172]]
postprocess_name: None
postprocess_dependencies: None
features: ['cg00112187', 'cg00151250', 'cg00359421', 'cg00513564', 'cg00532802', 'cg00574958', 'cg00668559', 'cg00782811', 'cg00835193', 'cg01055871', 'cg01101459', 'cg01360413', 'cg01554316', 'cg01936220', 'cg02004723', 'cg02079413', 'cg02229095', 'cg02300147', 'cg02307277', 'cg02571857', 'cg02650017', 'cg02949067', 'cg02997983', 'cg03604011', 'cg03776935', 'cg03810769', 'cg03868770', 'cg04051458', 'cg04105250', 'cg04305539']... [Total elements: 20000]
base_model_features: ['cg00112187', 'cg00151250', 'cg00359421', 'cg00513564', 'cg00532802', 'cg00574958', 'cg00668559', 'cg00782811', 'cg00835193', 'cg01055871', 'cg01101459', 'cg01360413', 'cg01554316', 'cg01936220', 'cg02004723', 'cg02079413', 'cg02229095', 'cg02300147', 'cg02307277', 'cg02571857', 'cg02650017', 'cg02949067', 'cg02997983', 'cg03604011', 'cg03776935', 'cg03810769', 'cg03868770', 'cg04051458', 'cg04105250', 'cg04305539']... [Total elements: 173]

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

base_model: LinearModel(
  (linear): Linear(in_features=20000, out_features=1, bias=True)
)

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

base_model.linear.weight: [-0.13975931704044342, 0.1523076742887497, 0.1058167889714241, -0.12104399502277374, 0.04729907587170601, -0.2645362913608551, -0.09107177704572678, 0.07626617699861526, -0.05811680853366852, -0.23175522685050964, 0.06270736455917358, 0.015027794055640697, 0.5146545171737671, -0.03559967875480652, 0.009041309356689453, 0.20905275642871857, 0.02438066340982914, 0.14796297252178192, 0.03653242066502571, 0.09093873202800751, -0.5942692160606384, -0.14602923393249512, 0.028370223939418793, 0.12909314036369324, 0.15818408131599426, 0.026989631354808807, -0.038831036537885666, 0.16809432208538055, 0.11093547195196152, -0.012085522525012493]... [Tensor of shape torch.Size([1, 173])]
base_model.linear.bias: tensor([-1.9499])

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

Basic test#

[14]:

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

[14]:

tensor([[ 0.2938],
        [ 0.8014],
        [-0.0384],
        [-0.2034],
        [-0.1848],
        [ 0.5177],
        [-0.1668],
        [ 0.3322],
        [-0.8489],
        [ 0.7232]], dtype=torch.float64, grad_fn=<AddmmBackward0>)

Save torch model#

[15]:

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

Clear directory#

[16]:

# 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
Deleted file: coefficients.xlsx
Deleted file: download.r
Deleted file: DunedinPACE.json
Contents