I am new to evolutionary genetics, and had problems in reading a paper about parallel selection in independent mice lines.
I could not agree with the interpretation that the authors made.
So let me first write what I understand about this paper,
and then my questions.
Multiple independent mice lines show similar phenotype due to similar selection pressure (parallel selection).
Body size is an example: there are lines with larger body weight throughout the world.
To identify the loci which control body weight, the authors searched for and identified genomic regions which had been under selection in independent lines with similar phenotype (Parallel Selected Regions or PSRs).
For example large-body lines tend to share genotypes at PRSs, but the genotype is not shared between big and small sized lines.
The background (other genomic region than PSRs) does not show this kind of association.
Then the authors searched for the signatures of selection at PSRs to convince us that the PSRs were actually under selection.
Generally, genomic regions under selection show 1. decreased heterozygosity, 2. excessive rare allele, 3. extended haplotypes, and 4. increased differentiation.
So, the authors calculated those 4 values near the PSRs and compared them with those of background (the other parts of genome).
The figure is below.
The blue line represents mice under no selection.
The yellow one represents mice under selection for 80 generations.
The blue one is mice under selection for 154 generations.
The authors wrote
"For all four statistics, we observed significant deviations in
PSRs compared to background loci."
"These signals became
stronger from generation 80 to 154 (Figures 3A–3D), suggesting
that ongoing parallel selective sweeps due to strong multilocus
selection produced the observed pattern of allele-sharing
across independent selection experiments."
This explanation did not convince me.
First, in panel B, C and D, the error bars are really big.
For example, the red line in panel B shows that the statistic of background is about 0.02 ± 0.017 whereas that of PSRs is 0.04 ± 0.025.
I do not understand why the difference between these two are significant.
Secondly, in A and B, at PSRs, the red line (154th generation) is below the yellow line (80th generation), which is below the blue line (control).
But this trend is also true in background: The background shows the same pattern of change according to generation numbers.
So, it seems to me that the signals (the difference between the statistic values at PSRs and that of background) did not change over generations.
What do you think of these questions.
Am I interpreting this paper in a wrong way?