The theory of evolution under threat - a new study

The random occurrence of mutations with respect to their
consequences is an axiom upon which much of biology and
evolutionary theory rests (https://go.nature.com/3g5fDDC).
This simple proposition has had profound effects on models
of evolution developed since the modern synthesis, shaping
how biologists have thought about and studied genetic diversity
over the past century. From this view, for example, the common
observation that genetic variants are found less often in
functionally constrained regions of the genome is believed to
be due solely to selection after random mutation. This paradigm
has been defended with both theoretical and practical arguments:
that selection on gene-level mutation rates cannot overcome
genetic drift; that previous evidence of non-random mutational
patterns relied on analyses in natural populations that were
confounded by the effects of natural selection; and that past
proposals of adaptive mutation bias have not been framed in the
context of potential mechanisms that could underpin such
non-random mutations
Yet, emerging discoveries in genome biology inspire a
reconsideration of classical views. It is now known that
nucleotide composition, epigenomic features and bias in
DNA repair can influence the likelihood that mutations occur
at different places across the genome. At the same time, we have
learned that specific gene regions and broad classes of genes,
including constitutively expressed and essential housekeeping
genes, can exist in distinct epigenomic states. This could in turn
provide opportunities for adaptive mutation biases to evolve by
coupling DNA repair with features enriched in constrained loci.
Indeed, evidence that DNA repair is targeted to genic regions and
active genes has been found. Here we synthesize these ideas by
investigating the causes, consequences and adaptive value of
mutation bias in the plant Arabidopsis thaliana.
De novo mutations in Arabidopsis .