Add Weidner aging clock (fixes #135)

[marcbal77]

Adds the Weidner blood-based aging clock (2014) to biolearn, implementing the 3-CpG age prediction model. This supersedes #135 and addresses all review feedback.

The Weidner clock uses DNA methylation at 3 CpG sites (ASPA, ITGA2B, PDE4C) to predict age from blood samples with ~5 year accuracy. Reference: Weidner et al., 2014

Changes:

• Added Weidner model definition with coefficients and linear transformation (intercept: 111.83)
• Added test data with verified expected outputs
• Fixed formatting and CSV issues
• Fixed documentation typo

Addresses review concerns from #135:

• Test data added for verification (biolearn/test/data/expected_model_outputs/Weidner.csv)
• Confirmed correct paper - Weidner is the first author

Closes #135, closes #170
cc: @chancenlaw @alexiszxcv (original PR authors), @sarudak (reviewer)

[sarudak]

The testset data was committed. Please remove and ensure that the large methylation file is cleaned out of the history.

[marcbal77]

Changes made:

1. I removed the large test data files from git history (testset_methylation_part0.csv and testset_metadata.csv). These are properly gitignored and get generated automatically when running make test - not sure how they got out of gitignore.

2. While reviewing the implementation, I noticed the coefficients from the original PR didn't match the published paper. I've corrected them to match the actual formula from Weidner et al. 2014 (https://doi.org/10.1186/gb-2014-15-2-r24):

   • The paper specifies: Predicted age = 38.0 - 26.4α - 23.7β + 164.7γ
   • Updated coefficients and intercept accordingly

The PR now only adds the 3 essential files (27 lines total) - the model definition, coefficients file, and expected test outputs.

cc: @chancenlaw @alexiszxcv - I preserved your original work, just fixed the coefficients to match the paper and cleaned up per review feedback + @sarudak

[sarudak]

Let's just move forward with what we have since we don't have any stronger validation for this.