- The first post from the brand new Nuts ‘n Bolts blog talks about hash kernels and how to use them to represent arbitrary input data in a format suitable for machine learning. There is a GitHub repo called hashkernel that demonstrates the approach. The tag line for the repo is great: A demonstration of how to use hash kernels for ridiculously unprincipled machine learning.
- This iPython notebook shows how to write a (greedy, not de Bruijn) genome assembler using tools available at the Pacific Biosciences GitHub repo. Titus Brown also has a repo showing how to implement a de Bruijn graph based ASCII assembler on top of Bloom filters.