epiTOC2

epiTOC2#

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:

[30]:

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

Instantiate model class#

[31]:

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

print_entire_class(pya.models.epiTOC2)

class epiTOC2(pyagingModel):
    def __init__(self):
        super().__init__()
        self.delta = None
        self.beta0 = None

    def preprocess(self, x):
        """
        Replace NaNs with zero; missing features should already be imputed via reference_values.
        """
        return torch.nan_to_num(x, nan=0.0)

    def forward(self, x):
        x = self.preprocess(x)

        device = x.device
        dtype = x.dtype

        delta = self.delta.to(device=device, dtype=dtype)
        beta0 = self.beta0.to(device=device, dtype=dtype)

        denom = delta * (1 - beta0)
        denom = torch.where(denom == 0, torch.ones_like(denom), denom)

        contrib = (x - beta0) / denom
        k = contrib.size(1)
        vals = 2.0 * torch.sum(contrib, dim=1) / k

        return self.postprocess(vals.unsqueeze(1))

    def postprocess(self, x):
        return x


[32]:

model = pya.models.epiTOC2()

Define clock metadata#

[33]:

model.metadata["clock_name"] = 'epitoc2'
model.metadata["data_type"] = 'methylation'
model.metadata["species"] = 'Homo sapiens'
model.metadata["year"] = 2020
model.metadata["approved_by_author"] = '⌛'
model.metadata["citation"] = "Teschendorff, Andrew E. \"A comparison of epigenetic mitotic-like clocks for cancer risk prediction.\" Genome Medicine 12.1 (2020): 56."
model.metadata["doi"] = "https://doi.org/10.1186/s13073-020-00752-3"
model.metadata["research_only"] = None
model.metadata["notes"] = "Stem cell division rate estimate using EpiTOC2."

Download clock dependencies#

Download coefficient file#

[34]:

coeff_url = "https://raw.githubusercontent.com/bio-learn/biolearn/master/biolearn/data/EpiTOC2.csv"
os.system(f"curl -L {coeff_url} -o EpiTOC2.csv")

[34]:

0

Load features#

From Excel file#

[35]:

df = pd.read_csv('EpiTOC2.csv', index_col=0)
df['feature'] = df.index.astype(str)
model.features = df['feature'].tolist()

Load weights into base model#

[36]:

#### From CSV file

[37]:

model.delta = torch.tensor(df['delta'].values, dtype=torch.float32).unsqueeze(0)
model.beta0 = torch.tensor(df['beta0'].values, dtype=torch.float32).unsqueeze(0)
model.base_model = None

[38]:

#### Linear model

Not used; computation happens in the model forward#

[39]:

model.reference_values = [-1]*len(model.features)

model.reference_values = [0.0] * len(model.features)

[40]:

model.preprocess_name = "mean"
model.preprocess_dependencies = None

[41]:

model.preprocess_name = "nan_to_zero"
model.preprocess_dependencies = None

model.postprocess_name = None model.postprocess_dependencies = None

[42]:

pya.utils.print_model_details(model)


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

training: True
metadata: {'approved_by_author': '⌛',
 'citation': 'Teschendorff, Andrew E. "A comparison of epigenetic mitotic-like '
             'clocks for cancer risk prediction." Genome Medicine 12.1 (2020): '
             '56.',
 'clock_name': 'epitoc2',
 'data_type': 'methylation',
 'doi': 'https://doi.org/10.1186/s13073-020-00752-3',
 'notes': 'Stem cell division rate estimate using EpiTOC2.',
 'research_only': None,
 'species': 'Homo sapiens',
 'version': None,
 'year': 2020}
reference_values: [-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]... [Total elements: 163]
preprocess_name: 'nan_to_zero'
preprocess_dependencies: None
postprocess_name: None
postprocess_dependencies: None
features: ['cg00043095', 'cg00347369', 'cg00397986', 'cg00466268', 'cg00884606', 'cg00916884', 'cg01435574', 'cg01537995', 'cg01587896', 'cg01699217', 'cg01783070', 'cg01830294', 'cg02150988', 'cg02186542', 'cg02266732', 'cg02631468', 'cg02726121', 'cg02796545', 'cg02964724', 'cg03045635', 'cg03111498', 'cg03140968', 'cg03181582', 'cg03430846', 'cg03450948', 'cg03603951', 'cg03874199', 'cg04188273', 'cg04408488', 'cg04431946']... [Total elements: 163]
base_model_features: None
base_model: None
delta: [4.999999873689376e-05, 4.999999873689376e-05, 4.999999873689376e-05, 9.999999747378752e-05, 9.999999747378752e-05, 4.999999873689376e-05, 4.999999873689376e-05, 4.999999873689376e-05, 9.999999747378752e-05, 4.999999873689376e-05, 4.999999873689376e-05, 0.0002500000118743628, 9.999999747378752e-05, 4.999999873689376e-05, 4.999999873689376e-05, 4.999999873689376e-05, 4.999999873689376e-05, 9.999999747378752e-05, 4.999999873689376e-05, 0.0002500000118743628, 9.999999747378752e-05, 4.999999873689376e-05, 4.999999873689376e-05, 4.999999873689376e-05, 9.999999747378752e-05, 4.999999873689376e-05, 4.999999873689376e-05, 9.999999747378752e-06, 4.999999873689376e-05, 4.999999873689376e-05]... [Tensor of shape torch.Size([1, 163])]
beta0: [0.019999999552965164, 0.05000000074505806, 0.05000000074505806, 0.05000000074505806, 0.009999999776482582, 0.05000000074505806, 0.05000000074505806, 0.05000000074505806, 0.029999999329447746, 0.05000000074505806, 0.03999999910593033, 0.0, 0.029999999329447746, 0.05000000074505806, 0.019999999552965164, 0.05000000074505806, 0.029999999329447746, 0.019999999552965164, 0.05000000074505806, 0.029999999329447746, 0.029999999329447746, 0.03999999910593033, 0.05000000074505806, 0.05000000074505806, 0.029999999329447746, 0.009999999776482582, 0.029999999329447746, 0.019999999552965164, 0.019999999552965164, 0.03999999910593033]... [Tensor of shape torch.Size([1, 163])]

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


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


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

Basic test#

[43]:

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

[43]:

tensor([[ 1562.6673],
        [-3824.6137],
        [  645.1144],
        [-1810.9748],
        [ -142.2699],
        [ 1630.9714],
        [ 9229.4758],
        [ 4666.7934],
        [ 5489.9210],
        [ 1256.8648]], dtype=torch.float64)

Save torch model#

[44]:

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

Clear directory#

[45]:

# 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: EpiTOC2.csv
Contents