StatQuest: MDS and PCoA

TL;DR

Learn how MDS and PCoA create plots based on distances among samples, leading to different clustering results compared to PCA.

Transcript

that quest is groovy just like a movie groove hello I'm Josh stormer and welcome to stat quest today we're going to be talking about multi-dimensional scaling MDS and principle coordinate analysis PCO a first of all if you don't have principal component analysis PCA down cold check out the stat quest PCA main ideas in five minutes principal compone... Read More

Key Insights

• ❓ MDS and PCoA focus on distances among samples, unlike PCA which focuses on correlations.
• 🙏 Different distance metrics can be used in MDS and PCoA like Euclidean, Manhattan, log fold changes, etc.
• 📈 The choice of distance metric in MDS/PCoA impacts how samples cluster in the final plot.
• ⚾ Similar fancy math is used in MDS, PCoA, and PCA, but the outcomes differ based on the focus on distances.
• 💯 MDS and PCoA output coordinates for a graph, variation percentages, and loading scores for variables' effects.
• ❓ Understanding the differences between PCA, MDS, and PCoA is crucial in data analysis and visualization.
• 📈 The selection of the best distance metric is crucial in MDS and PCoA and is considered part of the art of data science.

Questions & Answers

Q: What is the main difference between PCA and MDS/PCoA?

The main difference is that PCA focuses on correlations among samples, while MDS and PCoA focus on distances among samples, leading to different clustering results.

Q: How do MDS and PCoA calculate distances between samples?

MDS and PCoA use various distance metrics like Euclidean, log fold changes, Manhattan, and others to calculate distances between samples and create graphical representations.

Q: Why do biologists often use log fold changes to calculate distances in MDS/PCoA?

Biologists are interested in log fold changes among genes, making it a relevant metric for calculating distances in MDS and PCoA when studying biological samples.

Q: How do MDS and PCoA help in understanding the relationships between samples?

MDS and PCoA create 2D or 3D plots based on distances among samples, showing how closely or distantly related samples cluster together, aiding in visualizing relationships.

Summary & Key Takeaways

• MDS and PCoA create plots based on distances among samples, showing how closely samples cluster together.

• Different distance metrics like Euclidean, log fold changes, Manhattan, and more can be used in MDS and PCoA.

• While PCA focuses on correlations, MDS and PCoA focus on distances among samples to create graphical representations.