Table 1.

Published combinatorially complete fitness landscapes^a

System (assay)	Number of mutations	Number of genes	Largest combinatorially complete subset	Study’s main findings	Figure 1 panel	Citation
Avian lysozyme (melting temperature)	3	1	3	No selectively neutral pathway links the only two extant alleles	b	[25]
D. melanogaster visible mutant (productivity and male mating success)	5	5	5	Epistasis and sexual selection may attenuate genetic load in natural populations. Higher-order epistasis observed.	c	[39]
E. coli dihydrofolate reductase mutants (in vitro enzymatic activity)	5	1	3	Fitness landscape smoother than random; first study to offer quantitative definition of roughness	d	[42]
E. coli isopropyl malate dehydrogenase mutants (growth rate)	7	1	7	Essentially all epistasis for fitness arises in mapping from biochemistry to fitness	e	[26,34]
E. coli β-lactamase mutants (resistance against two antibiotics)	5	1	5	Sign epistasis constrains the number of selectively accessible mutational trajectories to highest-fitness allele; adaptive trajectories are rarely reversed when environment changes	f	[27,37••]
Solinaceae sequiterpine mutants (5-EA synthesis)	9	1	6	Rugged landscape in which alternate catalytic specificities are often mutationally nearby	g	[43]
A. niger visible mutations (growth rate)	8	8	5	Genetic recombination does little to speed adaptation; fitness landscapes have intermediate ruggedness	h	[22•,40]
P. falciparum dihydrofolate reductase mutants in E. coli (resistance against an antimalarial drug)	4	1	4	Clinical data consistent with evolutionary trajectory predicted from in vitro results	i	[28]
Mammalian glucocorticoid receptor mutants (cortisol binding)	4	1	4	Epistasis renders evolutionary trajectories selectively irreversible	j	[32]
P. falciparum dihydrofolate reductase mutants in S. cervisiae (resistance against two antimalarial drugs)	5	1	3	Landscapes not well correlated across environments	k	[33,38]
S. cerevisiae visible mutations (growth rate)	6	6	6	Epistasis is variable and genetic recombination does little to speed adaptation	l	[41]
HIV glycoprotein mutants (in vitro infectivity)	7	1	5	Common, strong epistasis. Higher-order effects noted	m	[56]
Metholobacterium extorquens beneficial mutations in novel metabolic pathway (growth rate)	4	4	4	Negative pairwise epistasis among beneficial mutations	n	[19••]
E. coli beneficial mutations (growth rate)	5	5	5	Negative pairwise epistasis among beneficial mutations	o	[30••]

^aSorted by year of publication.