For a novel and somewhat dispiriting theory of economic divergence, read A Farewell to Alms, published this year, by Gregory Clark of the University of California at Davis.
Under the read length distributions used in this study (see below), this resulted in, on average, approximately 0.11-0.21 variant positions per read length for low divergence read pools, and 1.2-2.3 1.2-2.3 positions per read length for high divariante read positions
Indel call completeness was lower on average than SNP call completeness for low divergence read pools, peaking at ca. 59% completeness for 16× coverage depth, but was nearly indistinguishable from SNP call completeness for high divergence read pools.
Importantly, the divergence levels (and misincorporation frequencies) employed in this study also capture the challenges associated with mapping reads harbouring multiple variant positions, as sequence reads from high divergence read pools averaged more than one variant position per read length.
In order to present a clearer view of trends at different levels of divergence, analyses were performed on high and low divergence read pools separately rather than considering divergence level directly as an independent variable.
This trend was not quite as pronounced at high divergence, as on average SNP recovery ranged from ca. 0% (0.1× coverage depth) to ca. 47% (16× coverage depth) for high divergence read pools.
Increasing average read length of a read pool also resulted in a greater decrease in average length of mapped reads across all damage levels and at both low and high divergence (Pr <<0.001 for linear regression), such that average lengths of mapped reads at 40/60/80 bp expected size were 40.05/59.99/79.44 and 40.03/59.92/79.16 bp for low and high divergence read pools, respectively.
Nonetheless, the overall lower success rate of read mapping in high divergence read pools resulted in an approximately three-fold decrease in the percent of reads mapped when read lengths were increased from 40 to 80 bp, as opposed to an approximately two-fold decrease observed in low divergence pools.
GC content, read length and damage all had significant impacts on the proportion of read pools able to be mapped back to their reference at both levels of divergence, while coverage depth was not significantly correlated for either low or high divergence read pool mapping (Table 1).
The rate of decrease in read mapping success was very similar for low and high divergence read pools (Table 1, Additional file 3).
The choice of the mapping software is highly critical and the reliability depends on the presence/absence of indels and the divergence between reads and the reference genome.
