Supporting Materials for “Spin models inferred from patient-derived viral sequence data faithfully describe HIV fitness landscapes”

Let ~ sn ∈ {0, 1}1xN be an arbitrary sequence1 in an equilibrium ensemble ofM sequences sampled according to Hint[~s] using the Metropolis Monte Carlo (MC) algorithm [1]. This ensemble is comprised of sequences from six independent realizations of 10 steps each, where each step constitutes an attempt to flip a single site. We eliminated the first 10 steps for equilibration, and recorded every i state such that mod(i, 5N) = 0. The sequences in the MC ensemble (also referred to as “equilibrium ensemble") capture the statistical properties of the intrinsic fitness landscape as encoded by Hint[~s]. The “mean sequence" of the MC ensemble is defined as,