Figure 3

Distribution of proteins within families constructed using two distinct family building methods: our currently employed domain composition based clustering versus the single-linkage BLASTP-based clustering method originally described.A: Frequency distribution of family sizes created by the two methods. B: Difference between the two methods evaluated at the protein levelon a per protein basis. The difference in family size between domain-based clustering and the single-linkage clustering method (DBC – SLC) was calculated for each protein that was included in a family using both methods. The histogram shows the total number of proteins found at each size difference displayed on the abscissa, binned at increments of 10.