Cell - 8 September 2016

Figure S5. Batch Effects in Fitness Measurements, Related toFigure 2
The first 6 panels show batch-batch fitness comparisons. Here we show the best replicate from each batch to highlight deviations in the best of our mea-
surements. Note that this is slightly different fromFigures 2C and 2D, where we do the batch-batch comparisons using the weighted average fitness across
replicates in the batch. Systematic divergences are at a scale of 1%–2% (per generation) for diploids, and larger for higher fitness lineages. The lastpanel shows
fitnesses for GPB2, PDE2, and diploid classes with the batches separated by black bars. We found that within replicate variation is in the 1%–2% range, while the
variation between batches can be as large as 5%. Systematic variation differs for different classes.