In this report, you will continue to analyze the same dataset as before: raw sequencing data of ancient DNA from the Neanderthal and Denisovan individuals of the Denisovan cave in the Altai mountains of Siberia. The question you should answer is: Does total coverage vary with nearby GC content, and if so, on what scale? Answer this for each individual and compare; you should also assess goodness of fit of your final model.
Here are my suggestions for how do do this:
Measure GC content as percentage of bases within \(\pm K\) that have true nucleotide G or C. (hint: running mean) Do this for several choices of \(K\), large and small, and include all of them in your model as potential predictors.
Don’t use the entire dataset: it will be too slow. Start out with a small, randomly chosen subset of the data, and then increase the size of the subset until you get tight estimates.
Look up poisson_log( ) in the Stan reference.
Note: You may assume that all bases in the dataset are consecutive along the genome. (This is mostly true.)