MammalianLifespan#

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

    def preprocess(self, x):
        return x

    def postprocess(self, x):
        """
        Applies an anti-log transformation.
        """
        return torch.exp(x)

[3]:
model = pya.models.MammalianLifespan()

Define clock metadata#

[4]:
model.metadata["clock_name"] = "mammalianlifespan"
model.metadata["data_type"] = "DNA methylation"  # Paper: The model is based on DNA methylation measurements.
model.metadata["species"] = "multiple species"  # Paper: The study samples are multi.
model.metadata["year"] = 2024
model.metadata["approved_by_author"] = "⌛"
model.metadata["citation"] = "Li, Caesar Z., et al. \"Epigenetic predictors of species maximum life span and other life-history traits in mammals.\" Science Advances 10.23 (2024): eadm7273."
model.metadata["doi"] = "https://doi.org/10.1126/sciadv.adm7273"
model.metadata["notes"] = "Pan-mammalian tissue-agnostic elastic-net predictor fitted to log species maximum life span from conserved CpG methylation; pyaging exponentiates the linear output to years."
model.metadata["research_only"] = None
model.metadata["tissue"] = ["multi-tissue"]  # Paper: These samples spanned 59 unique tissue types and originated from 348 distinct mammalian species across 25 taxonomic orders.
model.metadata["predicts"] = ["species maximum lifespan"]  # Paper: We will refer to the predicted maximum life span, expressed in log years, as either the epigenetic maximum life span or DNAm maximum life span.
model.metadata["training_target"] = ["species maximum lifespan"]  # Paper: We used three distinct penalized regression models to predict the log-transformed values of maximum life span, gestation time, and age at sexual maturity for each species.
model.metadata["unit"] = ["years"]  # Paper: We will refer to the predicted maximum life span, expressed in log years, as either the epigenetic maximum life span or DNAm maximum life span.
model.metadata["model_type"] = "elastic net regression"  # Paper: First, we used elastic net regression models to predict maximum life span using both CpG methylation data and taxonomic order indicators.
model.metadata["platform"] = ["Horvath MammalMethylChip40"]  # Paper: All data were generated using the mammalian methylation array (HorvathMammalMethylChip40), which provides high sequencing depth of highly conserved CpGs in mammals.
model.metadata["population"] = "multiple mammalian species"  # Paper: Leveraging our publicly accessible data from the Mammalian Methylation Consortium, we focused on highly conserved cytosine methylation profiles from n = 15,000 DNA samples. These samples spanned 59 unique tissue types and originated from 348 distinct mammalian species across 25 taxonomic orders.
model.metadata["journal"] = "Science Advances"
model.metadata["last_author"] = "Steve Horvath"
model.metadata["n_features"] = 152  # Paper: The official LifespanPredictor_40K_Li2021.csv contains 152 nonzero non-intercept CpG coefficients.
model.metadata["citations"] = 5
model.metadata["citations_date"] = "2026-07-05"

Download clock dependencies#

Download GitHub repository#

[5]:
github_url = "https://github.com/caeseriousli/MammalianMethylationPredictors.git"
github_folder_name = github_url.split('/')[-1].split('.')[0]
os.system(f"git clone {github_url}")
[5]:
0

Load features#

From CSV file#

[6]:
df = pd.read_csv('MammalianMethylationPredictors/Predictors/LifespanPredictor_40K_Li2021.csv')
df['feature'] = df['CpG']
df['coefficient'] = df['Coefficient']
df = df[df.Coefficient != 0]

model.features = df['feature'][1:].tolist()

Load weights into base model#

[7]:
weights = torch.tensor(df['coefficient'][1:].tolist()).unsqueeze(0)
intercept = torch.tensor([df['coefficient'][0]])

Linear model#

[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 = [0.5] * len(model.features)

Load preprocess and postprocess objects#

[10]:
model.preprocess_name = None
model.preprocess_dependencies = None
[11]:
model.postprocess_name = 'anti_log'
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': 'Li, Caesar Z., et al. "Epigenetic predictors of species maximum '
             'lifespan and other life history traits in mammals." bioRxiv '
             '(2023): 2023-11.',
 'clock_name': 'mammalianlifespan',
 'data_type': 'methylation',
 'doi': 'https://doi.org/10.1101/2023.11.02.565286',
 'notes': None,
 'research_only': None,
 'species': 'multi',
 'version': None,
 'year': 2023}
reference_values: [0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5]... [Total elements: 152]
preprocess_name: None
preprocess_dependencies: None
postprocess_name: 'anti_log'
postprocess_dependencies: None
features: ['cg00039845', 'cg00300233', 'cg00810217', 'cg01020408', 'cg01266508', 'cg01309159', 'cg01786675', 'cg02476543', 'cg02574410', 'cg02725055', 'cg02871478', 'cg03230916', 'cg03264110', 'cg03280886', 'cg03528345', 'cg03537184', 'cg03684591', 'cg03820088', 'cg04065686', 'cg04118146', 'cg04313551', 'cg04324237', 'cg04486940', 'cg04499301', 'cg04725401', 'cg04918691', 'cg04958799', 'cg05035746', 'cg05039938', 'cg05151611']... [Total elements: 152]
base_model_features: None

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

base_model: LinearModel(
  (linear): Linear(in_features=152, out_features=1, bias=True)
)

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

base_model.linear.weight: [0.04147933050990105, 0.4304673969745636, 0.07260479032993317, 0.10754379630088806, 0.05837688222527504, 0.21564741432666779, 0.22965361177921295, 0.0423436276614666, -0.6524762511253357, -0.3474166989326477, 0.05646226927638054, 0.046050943434238434, -0.03867235779762268, -0.14654619991779327, 0.29777470231056213, -0.22357487678527832, 0.19166646897792816, 0.015969855710864067, 0.12260589003562927, 0.07850058376789093, 0.11844679713249207, 0.1271219253540039, -0.3257320821285248, -0.0814819186925888, 0.1456708163022995, 0.037799667567014694, 0.19944973289966583, -0.14140872657299042, -0.0007183622219599783, 0.28238773345947266]... [Tensor of shape torch.Size([1, 152])]
base_model.linear.bias: tensor([-2.8285])

%==================================== 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([[1.8245e+00],
        [7.4422e-04],
        [1.6053e+00],
        [2.6872e-01],
        [3.1673e-03],
        [2.4528e+01],
        [5.4639e+00],
        [1.6672e-02],
        [1.2177e-01],
        [1.5087e-02]], dtype=torch.float64, grad_fn=<ExpBackward0>)

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 folder: MammalianMethylationPredictors