Skip to main content

Table 2 Hypotheses relating network constraint to evolutionary outcomes and results of hypothesis assessment using a node classification scheme in yeast

From: Aligning functional network constraint to evolutionary outcomes

Evolutionary outcome

Hypothesis (H)

Alternative Hypothesis (HA)

Results in this paper following assessment with hierarchical node classification scheme.

Speed of evolution

Indispensable or essential genes are more constrained and evolve slowly [43].

Functionally important and thus functionally constrained genes evolve slowly, independent of dispensability [23].

Highly expressed genes evolve slowest [15, 16].

HA: Functionally most constrained genes (H-nodes) have the lowest substitution ratios of all categories, and are most highly expressed, but have lower scores of evolutionary rewiring than P and I-nodes.

Speed of evolution

Central nodes have the highest number of edges; evolve very slowly because any change will lead to maladaptive pleiotropic effects - causing balancing selection through cost of complexity.

[38], [37], [20], [36]

Intermediate nodes evolve fastest as their higher number of edges allows for evolution through rewiring

[44, 45].

HA: Nodes with highest number of edges are intermediate to the network, evolve fast (high ɷ) and have a high score of rewiring (ɣ), indicating that the substitution rate of these genes may be associated with evolutionary rewiring events.

Speed of evolution

Nodes with a low number of edges evolve fastest due to higher degrees of freedom, which allows for genetic adaptations minimizing pleiotropic effects [46], [38]

H: Peripheral nodes evolve fast (high ɷ) and have a high score of rewiring (ɣ), indicating that the substitution rate of these genes may be associated with evolutionary rewiring events.

Convergent evolution

Nodes with a low number of edges should be the prime target of convergent evolution. Pleiotropic negative effects are expected to be low, and mutations in them can maximize adaptation [38].

Peripheral nodes have the highest degrees of freedom and thus divergence is more likely than convergence in them. Convergent evolution should instead be favored in nodes that allow for genetic variance, while having reduced degrees of freedom (I-nodes)

(This contribution).

HA: 21 out of 26 nodes with convergent evolution demonstrated in yeasts were classified as I- nodes by DFA, and five as P nodes. ɷ and CAI were similar to I-nodes, but none of these 26 nodes showed evidence of evolutionary rewiring.

Genic evolution

Adaptations can be characterized (either causative or correlative for the speciation process) by any number of divergent genes within the genome, whereas other genes are not associated with adaptation [47].

Only the complete phenotype is selected, the genic component is less important [10].

H: Different clusters of functionally similar nodes experience either higher, lower than expected or neutral rates of evolution across five species of yeast [41]. Causation or correlation to the speciation process not testable with data.