Embedded Files
Evidence for HIV as selective pressure in Lesotho?
Evidence for HIV as selective pressure in Lesotho?
The test that follows is a half-cooked and unfinished report from a dataset I analysed in 2021/2022 during my first postdoc in Parma. I never followed up on it and the analyses presented here are very rudimentary and basic. In principle I found no evidence.
Under the assumption that sampling was made at random, the percentage of people sampled from each age group in Lesotho and Namibia should yield similar country population demographic pyramids in the year 2009/2010. As far as I am aware, the sample collection process consisted in visiting villages in the Lowlands and Highlands of Lesotho and contacting the local elder or chief and explaining the project (permissions by the Lesotho and Namibia governments had been previously granted). There were no rewards involved for participants and it was completely voluntary. Participants could withdraw participation and consent at any moment, even after sample collection. Volunteers were only discouraged to participate if a known relative had already donated saliva. In fact I only found a few instances of kindred individuals in the genetic analyses.
Compared to Namibia, we identified an age group that appears under-represented in Lesotho. Potentially indicating an excess of mortality or emigration in this age group. This group was between the ages of 39-49 in 2009, in other words, born between the years 1960 and 1979 (Fig 1).
Figure 1: Population pyramids of Lesotho and Namibia based on the years of birth of the participants. Created from the dates recorded in the questionaire forms filled by participants during field work and saliva sample collection.
We investigated whether this could be due to the effects of the HIV epidemic of the 90’s. Based on public data (Fig 2), mortality rates spiked high in African countries such as Lesotho and Namibia in the early 90’s, but not so much in others, like Nigeria. Coincidentally, people born between 1960 and 1980 were young and likely to be more sexually active during the 90’s and therefore more exposed to the HIV virus.
Figure 2: Mortality rates over time in various countries. Screenshot sourced from https://databank.worldbank.org/, the actual sources are detailed in the figure.
Figure 3: Tentative population pyramids for Lesotho in 2009. Emigration rate in Lesotho by age group in 2013 (left) plus uncorrected (middle) and corrected (right) population pyramids of Lesotho in 2009.
The two pyramids are very different, in particular the lack of people born from the 1960s and onwards, and significantly so for people born in 1960-1980. However, since Lesotho experiences substantial labor emigration to South Africa, we corrected the Lesotho population pyramid for the expected missing percentage of people from each group according to the emigration rate reported for each age group based on an official report from 2013 by the Lesotho government (Fig 3).
After correcting for the emigration effect, a lack of people born between 1960 and 1980 is still observable in our sample. People of 40-44 years of age in 2009 (20-24 year-olds at the beginning of the 90’s) seem the most affected by a potential elevated mortality.
Figure 3: Tentative population pyramids for Lesotho in 2009. Emigration rate in Lesotho by age group in 2013 (left) plus uncorrected (middle) and corrected (right) population pyramids of Lesotho in 2009.
First exploratory analysis for possible HIV-related natural selection was checking whether the allele frequency of SNP rs9264942 in Chromosome 6 has changed dramatically over the last decades. Marker rs9264942 is linked to susceptibility to HIV-1 viremia and significant reductions of the viral load of HIV (REFERENCE). However no apparent great shift in allele frequency over time or across age groups is observable (Fig 4).
Figure 4: Allele frequency of rs9264942 over time and across age groups. Note that sample size in the youngest age group (born in the 90’s) is very small (n=3) in a total size of n=100.
Figure 5 might be somewhat confusing. We intended to identify the SNPs that had maximum allele frequency divergence between age groups. We defined three functional groups: born pre-60’s (n=50), born during the 60’s-70’s (n=26) and born post-70’s (n=24). The leftmost panel in Figure 5 focuses on the top divergent allele frequencies of 250 SNPs (out of 250k SNPs) when comparing the pre-60’s group and the inter-60-70’s group. We then ran an enrichment analysis for those SNPs but we did not find any oustanding association to genes related to immune responses.
Figure 5: Top 250 allele frequency divergence comparisons between age groups. Maximization of differences between pre-60’s and inter-60-70’s on the left, between pre-60’s and post-70’s in the center, between inter-60-70’s and post-70’s on the right.
Following this test, we attempted a simple Hardy-Weinberg equilibrium test grouping all individuals into one population. The distribution of values representing the difference between expected and observed heterozygosity (H) follows a normal distribution with the vast majority of values around zero (Fig 6 top left). However there are several outlier SNPs across chromosomes (Fig 6 top center). Chromosome 6 concentrates a high number of these outliers in a peak around the MHC region where HLA genes reside (Fig 6 top right). We confirmed this is not an artifact due to imputation bias by checking the imputation accuracy of single SNPs, which mostly resulted being above the 95% threshold. However this region always violates HWE since it is under balancing selection in all human populations. Therefore it is hard to conclude if there was further selection pressure acting on it in Lesotho without a baseline value or control cases to compare against.
We also found other SNPs out of Hardy-Weinberg equilibrium, for example in Chromosome 7 in the MUC3A gene (Fig 6 bottom), which is a gene known to be related and interacts with the HIV virus and the ability to bind to cells but the aforementioned issue remains.
Figure 6a: Histogram of the values of the difference between He and Ho, distribution of such values across chromosomes and in bp locations in chromosome 6.
Figure 6b: Breakdown by chromosome and position (bp) of how much each marker deviates from expected heterozygosity (Bottom).
Figure 7: prevalence of Eppstein-Barr (EBV/HHV-4) virus by age group in Namibia and Lesotho.
EBV infection can be associated with a significant liver damage in immune-compromised patients such as human immunodeficiency virus (HIV) positive patients (https://www.sciencedirect.com/topics/agricultural-and-biological-sciences/human-herpesvirus-4). These viral infections can accelerate HIV disease and multiplex real-time PCR can be used for the early detection (Sachithanandham et al 2014). Human herpesviruses (HHVs) have a particularly high prevalence in certain high-risk populations and cause increased morbidity and mortality in patients with acquired immunodeficiency syndrome (AIDS) (Ren et al 2019).
Page updated
Google Sites
Report abuse