23, Python and Me: Using Machine Learning in Python to Analyze Consumer Genomics Data
summary
Shows exactly what files look like and the Python stack used: pandas/Polars for large text files, scikit‑learn PCA and classifiers, and simple preprocessing to turn 23andMe genotypes into numeric features. Demonstrates working with ~900,000‑row raw outputs and a public cohort (Shing et al. subset of the 1000 Genomes: ~850 people, 250k variants) reduced to ~15,000 feature columns; PCA axes explain ~9–10% of variance, 23andMe covers ~0.3% of the genome, humans are ~98% identical. PCA plots ancestry gradients simulate personal genomes
AI-written summary · about
chapters (6)
- 0:00Python and genomics intro
- 4:10Genomics crash course
- 8:2223andMe raw data
- 17:13Machine learning with genomics
- 28:51Ancestry multiclass prediction
- 29:54Summary
related talks
30mBig data biology for pythonistas: getting in on the genomics revolution
20mAdvancing Epigenetic Predictors with Scalable Machine Learning
42mPython’s role in accelerating biomedical science
21mInteractive Machine Learning on Genomics Data with Spark Notebook
12mHow data science can make your genome actionable
51mSimplicity For Scale- Analyzing 15 Million DNA Samples With Python
04h 02mHands-on Session on Data Science for Beginners
50mBuilding Intelligent Mobile Apps using Machine Learning