Rapid Adaptive Increase of Amylase Gene Copy Number in Indigenous Andeans

Rapid Adaptive Increase of Amylase Gene Copy Number in Indigenous Andeans | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results Rapid Adaptive Increase of Amylase Gene Copy Number in Indigenous Andeans View ORCID Profile Kendra Scheer , View ORCID Profile Luane J. B. Landau , View ORCID Profile Kelsey Jorgensen , View ORCID Profile Charikleia Karageorgiou , Lindsey Siao , View ORCID Profile Can Alkan , Angelis M. Morales-Rivera , Christopher Osborne , View ORCID Profile Obed Garcia , View ORCID Profile Laurel Pearson , View ORCID Profile Melisa Kiyamu , View ORCID Profile María Rivera-Chira , View ORCID Profile Fabiola Leon-Velarde , View ORCID Profile Frank Lee , View ORCID Profile Tom Brutsaert , View ORCID Profile Abigail W. Bigham , View ORCID Profile Omer Gokcumen doi: https://doi.org/10.1101/2025.03.25.644684 Kendra Scheer 1 Department of Biological Sciences, University at Buffalo , Buffalo, NY, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Kendra Scheer Luane J. B. Landau 1 Department of Biological Sciences, University at Buffalo , Buffalo, NY, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Luane J. B. Landau Kelsey Jorgensen 2 Department of Anthropology, University of California , Los Angeles, CA, USA 3 Department of Anthropology, University of Kansas , Lawrence, KS Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Kelsey Jorgensen Charikleia Karageorgiou 1 Department of Biological Sciences, University at Buffalo , Buffalo, NY, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Charikleia Karageorgiou Lindsey Siao 1 Department of Biological Sciences, University at Buffalo , Buffalo, NY, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Can Alkan 9 Department of Computer Engineering, Bilkent University , Ankara, Turkey Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Can Alkan Angelis M. Morales-Rivera 10 Departamento de Biología dentro de las Ciencias Naturales, Universidad de Puerto Rico en Cay ey , Cayey, Puerto Rico Find this author on Google Scholar Find this author on PubMed Search for this author on this site Christopher Osborne 1 Department of Biological Sciences, University at Buffalo , Buffalo, NY, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site Obed Garcia 3 Department of Anthropology, University of Kansas , Lawrence, KS Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Obed Garcia Laurel Pearson 4 Department of Anthropology, The Pennsylvania State University, State College , PA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Laurel Pearson Melisa Kiyamu 5 Laboratorio de Fisiología del Transporte de Oxígeno y Adaptación a la Altura, Laboratorios de Investigación y Desarrollo, Facultad de Ciencias e Ingeniería, Universidad Peruana Cayetano Heredia , Lima, Peru Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Melisa Kiyamu María Rivera-Chira 6 Departamento de Ciencias Biológicas y Fisiológicas, Laboratorios de Investigación y Desarrollo, Facultad de Ciencias e Ingeniería, Universidad Peruana Cayetano Heredia , Lima, Perú 7 Centro de Investigación para el Desarrollo Integral y Sostenible, Universidad Peruana Cayetano Heredia , Lima, Peru Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for María Rivera-Chira Fabiola Leon-Velarde 5 Laboratorio de Fisiología del Transporte de Oxígeno y Adaptación a la Altura, Laboratorios de Investigación y Desarrollo, Facultad de Ciencias e Ingeniería, Universidad Peruana Cayetano Heredia , Lima, Peru Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Fabiola Leon-Velarde Frank Lee 11 Department of Pathology and Laboratory Medicine, University of Pennsylvania Perelman School of Medicine , Philadelphia, PA, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Frank Lee Tom Brutsaert 8 Department of Exercise Science, Syracuse University , Syracuse, NY, 13244 Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Tom Brutsaert Abigail W. Bigham 2 Department of Anthropology, University of California , Los Angeles, CA, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Abigail W. Bigham For correspondence: omergokc{at}buffalo.edu Omer Gokcumen 1 Department of Biological Sciences, University at Buffalo , Buffalo, NY, USA Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Omer Gokcumen For correspondence: omergokc{at}buffalo.edu Abstract Full Text Info/History Metrics Supplementary material Data/Code Preview PDF Summary/Abstract The salivary amylase gene ( AMY1 ) exhibits remarkable copy number variation linked to dietary shifts in human evolution. While global studies highlight its structural complexity and association with starch-rich diets, localized selection patterns remain under explored. Here, we analyzed AMY1 copy number in 3,723 individuals from 85 populations, revealing that Indigenous Peruvian Andean populations possess the highest AMY1 copy number globally. A genome-wide analysis showed significantly higher amylase copy numbers in Peruvian Andean genomes compared to closely related populations. Further, we identified positive selection (selection coefficient of 0.0124, log likelihood ratio of 11.1543) at the nucleotide level on a haplotype harboring at least five haploid AMY1 copies, with a Peruvian Andean-specific expansion coinciding with potato domestication (∼6–10 kya). Using ultra-long-read sequencing, we demonstrated that previously described recombination-based mutational mechanisms drive the formation of high-copy AMY1 haplotypes observed in Andean population. Our study provides a framework for investigating structurally complex loci and their role in human dietary adaptation. Introduction The amylase enzyme plays a central role in the modern human diet. Its main function is to digest starch, a cornerstone of most contemporary diets, into simple sugars 1 . Among humans, salivary amylase ( AMY1 ) copy number variation demonstrates exceptional diversity, with individuals carrying between 2 and 20 copies per diploid genome, potentially affecting salivary amylase dosage and downstream starch hydrolysis 2 . Early duplications of AMY1 in hominin evolution facilitated further copy number expansions in humans 3 , which some studies have linked to agricultural diets 2 , 4 . A similar trend has been observed in other species consuming broad range diets 5 . For example, as they began cohabiting with humans, dogs experienced an increase in amylase gene copy number, especially in breeds living with agricultural communities 5 – 7 . The rise in amylase copy number in dogs likely led to a higher pancreatic amylase dosage and is associated with amylase content in saliva. Further, human amylase copy number variation has been associated with metabolic traits, including the composition of the gastrointestinal microbiome, which may shape metabolic and oral health 8 , 9 . Despite these exciting associations, the structural complexity of the AMY1 locus, characterized by recurrent duplications, deletions, and inversions, has complicated efforts to connect gene dosage to local dietary evolutionary pressures in extant human populations. The Andean Altiplano, where potato domestication began approximately 6,000–10,000 years ago 10 , 11 , offers a compelling model for investigating the adaptive significance of AMY1 copy number variation in extant human populations. Potatoes, a dietary staple in this region, are among the highest glycemic index foods, meaning that their consumption can cause spikes in blood glucose levels far more easily than other carbohydrates 12 . For thousands of years, Andeans have relied heavily on the nutrition provided by this staple, which still constitutes over 50% of daily caloric intake in the modern Andean diet 13 – 15 . Additionally, prior work has already demonstrated selection on the digestive enzyme, MGAM , in relation to the potato starch digestion process in Andeans 16 , 17 . Thus, potatoes may have played a significant role in shaping the evolution of genes involved in carbohydrate metabolism among Andeans. Here, we accurately genotyped AMY1 copy numbers in a cohort of Peruvian Andean individuals, combining population genomic, haplotypic, and structural analyses. By integrating ultra-long-read sequencing with genome-wide selection tests, we reveal a signature of local adaptation affecting the AMY1 copy number variation likely as a response to the starch-rich diet in Andean populations. These findings contribute to our understanding of local dietary adaptations and the evolutionary significance of copy number variation in complex genomic loci. Results Peruvians and Akimel O’odham/Pima have the highest AMY1 copy numbers among populations tested We calculated gene copy number variation from 3,723 individuals across 85 global populations using two read-depth-based genotyping approaches ( Table S1, see Methods) . These counts were validated with Digital Droplet PCR for 101 results and 33 previously published high-quality amylase haplotypes with r 2 values of 0.94 (p = 6.01 x 10 - 11 ) and 0.98 (p = 1.81 x 10 - 25 ), respectively 3 . We verified previous observations that Peruvians in Lima, Peru (PEL) possess strikingly elevated AMY1 copy numbers compared to other global populations ( Figure 1A ) 2 . In fact, PEL display one of the highest median copy numbers (PEL median = 10, n=85) along with the Pima population (preferably known as Akimel O’odham, median = 10, n=14) among all 85 populations (worldwide median = 7) tested in this study. Download figure Open in new tab Figure 1. ( A ) Boxplots of AMY1 diploid copy numbers across all populations, ranked by median. Populations are compiled using samples from the 1000 Genomes dataset, the HGDP dataset, individuals from Tibet and the Middle East, as well as Quechua and Maya individuals we have provided (see Methods). American populations have been split into gold (high copy number populations) and purple (low copy number populations). A black dotted line indicates the overall median (7). ( B ) Geographic distribution of all populations with circle size indicating median AMY1 copy number (per diploid cell) and colors representing regions. Population counts are found in Table S2 ( C ) Zoom in on American populations from panel B with populations demonstrating a higher proportion in AMY1 genes labeled in gold. ( D ) The histogram of the AMY1 copy numbers of the American populations from panel C, the three previously highlighted indigenous populations show an increase in the AMY1 copy number. (E) Bar plots displaying the fraction of different ancestry proportions for populations tested in this study. The Maya from Chiapas population contains the least amount of non-American ancestry while MXL contains the most. Prior work has demonstrated differing continent level trends regarding AMY1 copy number 2 , 18 . To assess the AMY1 copy number pattern of variation among populations within each continent, we performed a Kruskal-Wallis (KW) test ( Figure S1A ). This analysis revealed significant population structure in the Americas (p < 2.2e-16), while other continents showed relatively homogeneous AMY1 distributions ( Figure 1B , 1C , & 1D ). This high level of population structure is driven by PEL and Pima possessing the highest median copy numbers among all 85 populations tested, while other American populations exhibit some of the lowest median copy numbers ( Figure 1A ) . When we remove both the Peruvian and Pima populations, this population structuring is lost (KW p = 0.4) ( Figure S1B ), indicating that the observed signal is primarily driven by the high AMY1 copy numbers in these two populations. Due to the known European and African admixture in the American populations in the 1000 Genomes Project, we tested if this difference of AMY1 copy number among populations can be explained by varying amounts of African and European admixture. We therefore repeated our analysis in an independent cohort of Quechua speakers from Lima and Cerro de Pasco, Peru (n = 81) and in a control cohort of Mexican Maya speakers (Tzeltal, Tzotzil, and Ch’ol) from Palenque, Chiapas, Mexico (n = 48) ( Table S1 ) and likewise found exceptionally high AMY1 copy numbers in the Quechua cohort ( Figure 1A-D ). All of these individuals have high amounts of Indigenous American ancestry ( Figure 1E ; Tables S2-S3 ). We found that the Quechua population had AMY1 copy numbers comparable to the PEL population (both medians = 10, range = 4-20 and 2-19, respectively), while the Maya population had AMY1 copy numbers (median = 6, range = 2-13) that did not differ significantly from other American populations from the 1000 Genomes and Human Genome Diversity Project (HGDP) datasets. We note that the individual from the PEL population that had the lowest AMY1 copy number (2) had 100% African ancestry in the locus ( Table S4 ). Of the two Quechua individuals who carried just four AMY1 copies, one showed 100% European ancestry at the amylase locus and was therefore excluded from subsequent analyses ( Table S5 ). Overall, we conclude that the observed high levels of AMY1 copy number variation among American populations cannot be explained solely due to the percentage of Indigenous American ancestry ( Figure 1D & 1E ). AMY1 copy number is a genome-wide outlier in Peruvians We hypothesized that selection, perhaps in response to a starch-rich potato heavy diet, drove the high AMY1 copy number in Peruvian populations. Given that the indigenous Andean population is descended from populations that have undergone several bottle necklaces (including both the crossing of the Bering Strait and the entrance to South America) which may lead to population differentiation, we conducted genome-wide analysis to rule out neutral, demographic forces. First, we conducted a genome-wide VST outlier analysis, an FST analog designed for copy number variation 19 . We prioritized the comparison between Quechua and Maya individuals due to their relatively higher Indigenous American ancestry compared to 1000 Genomes populations ( Figure 1E ). Specifically, we compared gene copy number variants from genes that had a greater than 1 total copy number difference when rounded to an integer. We calculated these results using the program mrCaNaVaR between Peruvian Quechua individuals from Cerro de Pasco (n = 46) and Maya individuals from Palenque (n = 48) ( Figure 2A , Table S6 ). Because mrCaNaVaR cannot distinguish among closely related paralogs, we applied a second copy-number caller to determine whether the observed variation occurred in AMY1 or AMY2 . This analysis confirmed that the increase was specific to AMY1 , as AMY2A and AMY2B showed no significant variation across American populations ( Figures S2 & S3 ). AMY1 ranked among the top three most differentiated genes genome-wide in copy number, placing in the top ∼0.07% of all genes tested ( Table S7 ). Download figure Open in new tab Figure 2: ( A ) Scatterplot showing the relationship between Vst and median gene copy number difference in all protein-coding genes in Quechua from Cerro de Pasco and Maya from Palenque populations for genes that have a total copy number difference greater than 1. Genes that have zero difference between the populations (both Vst and the Median difference = 0) are not plotted. The AMY1 and AMY2 , are represented by a single dot that has been pointed out. Darker shading is associated with greater Vst and median copy number difference values. ( B ) Scatterplot showing the relationship between Vst and median gene copy number difference in all protein-coding genes in Peruvians in Lima and Mexicans in Los Angeles populations for genes that have a total copy number difference greater than 1. To further validate our findings, we replicated this analysis using 1000 Genomes data, comparing PEL and Mexicans from Los Angeles (MXL) ( Figure 2B , Table S6 ). This analysis corroborated our results, with AMY1 copy number variation ranking among the highest median differences observed for common copy number variable genes in the top seven genes (top ∼1.2% Table S7 ). The slight differences in the two comparisons may be partially influenced by increased admixture in the PEL and MXL populations. Nevertheless, population differentiation tests across two independent cohorts consistently showed that variation in AMY1 copy number exceeds genome-wide expectations, encouraging us to further investigate the hypothesis that high copy numbers in the Indigenous Andean population may result from a selective sweep. Identifying haplotypes harboring high AMY1 copy numbers in Peruvian Andean populations To understand the haplotype structure associated with high AMY1 copy numbers, we analyzed linkage disequilibrium (LD) using the Kruskal-Wallis test at the amylase locus and its flanking regions (hg38 Chr1:103,348,464–103,830,994) across populations in the 1000 Genomes Project and compared these patterns to those observed in the Quechua and Maya populations ( Figure S4; Table S8 ). We found that in most populations, flanking single nucleotide variants (SNVs) were not significantly correlated with AMY1 copy number. However, we identified 81 SNVs that showed significant correlation with AMY1 copy number and were shared between East Asian and Andean/Maya populations, suggesting the presence of a shared ancestral haplotype ( Table S9 ). These SNVs also were in long-range, high LD (r 2 ) with each other in the Quechua population ( Figure S5, see Github ). Further, we identified 7 region-specific SNVs that appear to have evolved on this ancestral haplotype in Indigenous American populations as well as 17 SNVs private to the Andean populations ( Table S9 ). These variants were also in strong LD with one another and weaker LD with the ancestral haplotype due to their later addition ( Figure S5 ). The combination of these SNVs form a population-specific haplotype block in Andeans that extends further on each side of the amylase locus than any other observed among the populations analyzed, and this haplotype was significantly correlated with high AMY1 copy numbers ( Figure S4 ). We hypothesized that if the increased LD at the amylase locus in the indigenous Andeans reflects a recent selective sweep, then the same SNVs that correlate with AMY1 copy number should exhibit increased FST between Peruvian Quechua and Maya populations, mirroring the trends observed in our VST outlier analysis. Consistent with these expectations, we found that the SNVs tagging the AMY1 copy numbers show significantly high FST values (ex. rs58677276 = 0.311 and rs143597860 = 0.199 are both above 99 th percentile) compared to the rest of chromosome 1 ( Figure 3A , see Github ). Download figure Open in new tab Figure 3: ( A ) Chromosome 1 SNVs in Quechua from Peru (top) and Maya from Chiapas (bottom) with Indigenous American ancestry in the amylase region ( Table S4 ) in the alternative allele. Y-axis: directional FST with SNVs that have a higher frequency in the Quechua population above the dotted axis and SNVs with a higher frequency in the Maya population below the dotted y-axis. The 99th percentile threshold is marked. X-axis: -log10(p-values) from a Kruskal-Wallis test comparing diploid AMY1 copy numbers among genotype classes. SNVs to the right of the axis are associated with an increase in AMY1 copy number and those to the left are associated with a decrease in AMY1 copy number. The vertical dashed line marks the Bonferroni-corrected significance threshold (α = 0.01); points further from the origin are significant after correction. Colors indicate the relationship between these statistics. Three groups of SNVs that seem to be associated with AMY1 copy number are highlighted. ( B ) rs58677276 which we have highlighted as the example SNV for group 2. Pairwise Wilcoxon values are labeled. All SNVs in this group are negatively associated with the AMY1 copy number and primarily appear in the Maya population. ( C ) Frequency of rs58677276 in American populations. ( D ) rs143597860 which we have highlighted as the example SNV for group 3. Pairwise Wilcoxon values are labeled. SNVs in this group are associated with an increase in AMY1 copy number. ( E ) rs143597860 frequency in American populations. When jointly considering population differentiation and correlation with AMY1 copy number, we identified three SNV clusters corresponding to distinct haplotypes that drive the observed differentiation between the Maya and Quechua at the amylase locus ( Figure 3A ). The first group comprises SNVs where the reference allele was associated with higher AMY1 copy number. The reference allele of an example SNV from this cluster, rs1930184, segregates at ∼74% in the Quechua and is also common in East Asian populations (∼60-68%), likely contributing to the higher AMY1 copy number observed in both groups. In Maya, the reference allele was found in only ∼42% of the population, contributing to the high FST. The SNVs in this cluster were the same aforementioned SNVs that share high LD between the EAS and Andean/Maya populations. A phylogeny of the SNVs in the flanking regions of the amylase locus displayed a high copy number clade that can be tagged by rs1930184 within the East Asian populations in which most of the high copy number haplotypes in Andeans and Maya were all nested ( Figure S6 ). The second group, here tagged by rs58677276, contains SNVs where the alternative allele was associated with lower AMY1 copy number ( Figure 3B ). The alternative allele for rs58677276 was present at frequency of ∼26% in the Maya and completely absent in Quechua ( Figure 3C ) though other SNVs in this cluster were present at frequencies of ∼1-2% in Quechua. The third cluster, tagged here by rs143597860, was specific to the Americas and includes SNVs where the alternative alleles were associated with high diploid AMY1 copy numbers ( Figure 3D ). Rs143597860 segregated in its alternative state at ∼46% in Quechua but only at ∼14% in Maya ( Figure 3E ) and had the highest Pearson correlation coefficient value between a SNV and AMY1 copy number (Pearson correlation 0.71, p-value = 3.3×10 - 13 ) ( Table S10 ). We hypothesize that this population-specific haplotype, tagged by rs143597860 and harboring five or more AMY1 copies, provides evidence that its frequency has increased in Indigenous Andean populations through a soft selective sweep. Selection on standing variation best explains the high allele frequency of high AMY1 copy number harboring haplotypes in indigenous Andeans To assess whether demographic bottlenecks alone could account for the observed allele frequency differentiation of high AMY1 copy number haplotypes, we conducted neutral simulations using msprime (see Methods ). Using published divergence times and effective population sizes, along with our own estimates via Relate ( Figure S7 ), we were unable to replicate the observed allele frequency of rs143597860 ( Table S11 ). Specifically, across 3×10 9 simulated base pairs (roughly the size of the human genome) we could not replicate a scenario where any mutated SNPs achieved a frequency of greater than 45% in Andeans and less than 13% in Maya under the estimated parameters. To further test whether higher genetic drift could account for the observed frequencies, we performed 192 additional simulations under progressively reduced (and increasingly unrealistic) effective population sizes for the Quechua and Maya populations ( Figure S8 , Table S11 ). Only under highly specific and implausible conditions, such as a Quechua Ne of 250 and a Maya starting Ne of 1,000 with a final Ne of 500, did the simulations yield allele frequencies approaching the empirical data. These findings suggest that drift alone is unlikely to explain the elevated frequency of this allele in Indigenous Andeans, pointing instead to a potential role for positive selection. Our observation that SNVs that are in unusually high correlation with AMY1 copy numbers exhibit significant population differentiation in Quechua suggests that LD patterns for these SNVs should differ markedly between Andean Peruvians and the closely related Maya population. To test this empirically, we calculated cross-population extended haplotype homozygosity (XP-EHH) between Quechua and Maya genomes 20 . As a sanity check for the accuracy of this test, we verified that the EGLN1 locus, a known target of high-altitude selection in Andean populations 21 , also displayed significantly high XP-EHH values (>4, >99 th percentile, Figure S9, see Github) , replicating previous studies and supporting the sensitivity of our analysis. We found that multiple SNVs flanking the AMY1 locus, including SNV rs143597860, also displayed chromosome-wide significant XP-EHH values more significant that what was observed for EGNL1 locus (>6, >99 th percentile Figure 4A , Figure S9, see Github ). This result is conservative, as most genetic maps are based on SNV data and exclude recombination events within the AMY1 locus due to segmental duplications that evolve more rapidly than SNVs. Moreover, the significant XP-EHH block flanks the AMY1 locus on both sides, containing multiple linked SNVs associated with AMY1 copy number. Download figure Open in new tab Figure 4: ( A ) XP-EHH between Quechua and Maya samples in the amylase locus and flanking regions along with regional recombination rates. Brown squares are AMY2B and AMY2A while orange squares are AMY1A , AMY1B , and AMY1C . Chromosome-wide significant values are in red. Positive XP-EHH values have longer haplotypes in Quechua; negative XP-EHH values have longer haplotypes in Maya. ( B ) Temporal trajectories of the derived allele rs143597860 in Quechua and Maya populations. Solid lines trace the maximum-a-posteriori (MAP) estimate of derived-allele frequency through time for each population. Shaded regions mark every frequency bin with non-zero posterior probability at each generation (i.e., the full posterior support and not a fixed-probability credible band). Dashed vertical bars denote the archaeologically inferred interval for potato domestication (6,000–10,000 BP). ( C ) A phylogeny of the haploid amylase and flanking regions. Bars are colored by population and shaded by AMY1 copy number, with the presence of rs58677276 ( Figure 3B , C ) and rs143597860 ( Figure 3D , E ) labeled. All Quechua individuals with ≥ 14 AMY1 copies (exceeding the max Maya copy numbers) have shared haplotypes linked by a dotted line. To further evaluate the putative soft selective sweep, we conducted additional tests to ensure that our observations were not confounded by admixture or drift. First, we applied the Population Branch Statistic (PBS) 22 , which builds on FST by incorporating an outgroup and is particularly useful for detecting selection in admixed populations. Using the FIN population as an outgroup and comparing PEL and MXL populations, we found that the tag SNV rs143597860 had one of the highest PBS values among all SNVs analyzed (top 0.04%), placing it well above the 99 th percentile ( Figure S10, see Github ). In parallel, we used Ohana 23 , which explicitly models admixture components across the genome and identifies SNPs with allele frequencies that deviate from genome-wide expectations. Running Ohana on the Quechua genomes using the Iberians in Spain, (IBS) and Yoruba in Ibadan, Nigeria (YRI) populations as comparisons, we found that rs143597860 again ranked in the top 0.01% of SNVs analyzed ( Figure S10, see Github ). Together, these selection scans consistently reveal elevated population differentiation and extended haplotype homozygosity at rs143597860, providing robust, multi-method evidence for recent positive selection on this haplotype in Peruvian Andeans. The putative soft sweep coincides with potato domestication in the Andes The Atlas of Variant Age with a 25-year generation time, dates rs143597860 as 7,068 years old based on a recombination clock, 27,368 years old based on a mutation clock, and 17,945 years old with a joint clock estimate, coinciding with early migrations to the Americas around 13,000-23,000 years ago 24 , 25 , based on 1000 Genomes and SGDP (Simons Genome Diversity Project) ( Figure S11 ) 26 – 28 . As this SNV segregates at a frequency of 3.7% in populations from Central Asia Siberia, rs143597860 likely existed among the ancestral population to both Maya and Quechua, increasing in frequency among Andeans in a manner concordant with selection on standing variation due to the hypothesized selective sweep on the high amylase copy numbers. To explore the evolutionary trajectory of rs143597860 and the likelihood of selection, we leveraged Relate to infer approximate ancestral recombination graphs and CLUES to estimate allele frequency trajectories over time, including the strength and likelihood of selection 29 , 30 . These analyses revealed that the allele G, the alternate allele of the putatively selected SNV rs143597860, segregated at very low frequencies (0.7-2.7%) in ancestral populations prior to the divergence of Quechua and Maya populations, before rising in frequency after ∼10,000 years ago ( Figure 4B & S12, see Github ). In the Maya, the allele rose modestly, slightly exceeding 10% frequency in the last 10,000 years. However, the low log-likelihood ratio (LLR) score (1.53) was statistically indistinguishable from neutrality and any evidence for selection on this locus in the Maya was inconclusive. In contrast, the rs143597860 trajectory in the Quechua population showed a striking increase from ∼2% to ∼45%, best explained by a selection coefficient of 0.0124 with a log-likelihood ratio of 11.15. This result strongly supports the action of a selective sweep. Notably, the timing of this allele frequency increase coincides with the period of potato domestication in the Andean region (6,000–10,000 BP), raising the intriguing possibility that dietary shifts may have contributed to the selective pressures acting on this locus. For comparison, the EGLN1 variant rs1769792, previously identified as a top target of selection at high altitude 31 , 32 has a selection coefficient (S) of 0.005 and an LLR score of 8.08 ( Figure S13 ). To put the aforementioned SNVs in context and to visualize the amylase haplotypic variation, we constructed a phylogenetic tree using phased SNVs from Quechua and Maya individuals encompassing the amylase locus and flanking regions (hg38: Chr1: 103,348,464-103,830,994) ( Figure 4C , see Github ). On this phylogeny, we superimposed the diploid AMY1 copy number and the presence or absence of the three high FST SNVs described above. We found that rs58677276 appeared mainly in a clade (corresponding to the 2 nd cluster in Figure 3A ) harboring diverse haplotypes that were predominantly composed of Maya individuals (25/27 haploid genomes) with low AMY1 copy numbers (median = 4). This phenomenon may have been caused by a deletion event triggering the loss of AMY1A and 1B , the core mutational unit described previously 3 , 18 , and resulting in a static single AMY1 copy number, which may be present in this clade’s haploid genomes. More relevant to the proposed selective sweep, we identified a clade that predominantly carried the alternative allele for rs143597860 (corresponding to the 3 rd cluster in Figure 3A and the highlighted high copy number clade in Figure 4C ), associated with the selected haplotype. This clade was largely Quechua (67/77 haploid genomes) and exhibited high AMY1 copy numbers (median = 12), accounting for over half (64 of 116) of all haploid genomes in our dataset who have diploid AMY1 copy numbers ≥ 10. When expanding the phylogeny to include EAS populations, we found this clade was nested within the high copy number clade tagged by rs1930184 ( Figure S6 ). Structural haplotypes of Peruvian populations The structural haplotypic variation of the amylase locus in Peruvians is not well-characterized. Previous studies showed that AMY1 copy number variation involves non-allelic homologous recombination (NAHR) of a mutational unit comprising AMY1A and inverted AMY1B ( Figure S14 ) 3 . To address whether the unusually high AMY1 copy number haplotypes in Peruvian Andeans were due to the same mutational mechanism, we analyzed data from 40 existing haplotype assemblies, including 10 Peruvian samples ( Figure 5A ; Table S12 ) 2 , 3 . The existing assemblies allowed us to construct a pangenome of the amylase locus representing 33 distinct structural haplotypes ( Table S12 , see Github ). Among these, 14 haplotypes were present in Peruvians, yet all of them also occurred in other global populations ( Figure 5B ). Download figure Open in new tab Figure 5: Peruvian populations have uniquely long haplotypes while maintaining the same duplication mechanisms. ( A ) Overview of the analytical approach used to identify and characterize Peruvian haplotypes and their underlying mutational mechanisms. ( B ) Haplotypes found in Peruvian individuals. Known haplotype structures reported in previous studies are shown alongside newly inferred Peruvian haplotypes identified in this study. The asterisk corresponds to the manually curated H7 haplotype present in Peruvians. ( C ) Scatter plot illustrating the relationship between mean read depth (y-axis) and standard deviation (x-axis) for individual HG02106’s reads mapped against all known haplotypes. The highlighted data points represent the best-fitting mappings for HG02106’s two haplotypes (H7), characterized by their lower standard deviation values. ( D ) The upper panel on the left shows the per-nucleotide read depth for HG02106 when aligned to the H9 haplotype. In contrast, the lower panel shows the alignment to the inferred H7 haplotype, demonstrating higher median read depth and reduced standard deviation, suggesting a better haplotype fit. On the right, the read depth when aligned to an H3 haplotype, shows the ratio between the region with invariable copy number of AMY2B and the region comprising the repeat AMY1A / AMY1B . To further investigate, we generated long-read sequencing data from six PEL individuals representing a range of copy number alleles, including those carrying the highest AMY1 copy numbers Figure 5A ). These six PEL samples were children in independent trios with previously sequenced genomes ( Supplementary Methods) . Specifically, we generated Oxford Nanopore long-read sequencing data ranging from 14–45× coverage ( Supplementary Table S13 ), achieving an N50 of ∼20 kb. For two samples, we additionally obtained ultra-long reads with an N50 of ∼50 kb. Using this dataset, we constructed phased contigs through local assembly tools ( Figure S14, Supplementary Methods, Table S13 ). We analyzed the pangenome and individual contigs for sequence signatures indicative of previously characterized microhomology-mediated break-induced replication (MMBIR) events, which have been implicated in rare occurrences of amylase duplications 3 . No such signatures were detected in the Peruvian samples, either within the pangenome or the contigs ( Supplementary Methods ), suggesting that observed structural variation at the AMY1 locus was not driven by MMBIR mechanisms in this population. This conclusion is further supported by the absence of significant variation in AMY2A and AMY2B copy numbers in Peruvian Andeans ( Figure S2 & S3 ), which are shown to be typically affected by MMBIR. Next, we examined the above-mentioned 40 assembled haplotypes to confirm that the observed AMY1 copy number variation in the Peruvian samples could be explained by the recently described non-allelic homologous recombination (NAHR) mechanisms involving both AMY1A and AMY1B ( Figure 5B ). We competitively mapped the newly sequenced phased long reads from six PEL individuals (12 phased samples) onto the 40 assembled haplotypes. For each phased sample, we calculated the mean read depth and standard deviation of read depth across all assembled haplotypes to identify the best fit (e.g., Figure 5C , Supplementary Methods ). We assumed that the best-matching haplotype would exhibit the most uniform coverage across assemblies (i.e., lowest standard deviation) and an appropriate mean read depth (see Supplementary Methods ). These assignments were further validated by copy number estimates derived from GeneToCN and Droplet Digital PCR analyses ( Table S1) 3 . Overall, we successfully assigned 8 out of 12 newly sequenced phased samples to known reference haplotypes ( Table S13, S15 ). Two phased samples (HG01975 and HG02106), with an expected 14 diploid copies each, did not match the expected copy numbers by the read-depth approach ( Supplementary Methods ). We manually inspected the mapping to known haplotypes and found that the alignments to other known 3-, 5-, and 9-copy haplotypes were not uniform. Furthermore, despite being the best match, we observed apparent gaps in the mapping even for the 9-copy previously resolved Peruvian haplotype ( Figure 5D ). This led us to hypothesize that, despite structural similarities, the Peruvian haplotypes may differ in local sequence variation. To test this, we generated an estimated 7-copy haplotype by deleting one mutational unit from the 9-copy Peruvian haplotype ( Figure S16 ). This adjustment significantly improved the mapping ( Table S15 ), providing the best fit for these phased samples ( Figure 5C ), and allowed us to conclude that this individual may carry two 7-copy haplotypes ( Figure 5D ), and observed exclusively, with the exception of two African individuals, in East Asian and American populations to date 33 . Together, these results suggest that the majority of high AMY1 copy number haplotypes observed in Peruvians already existed in global populations and have evolved through the same mutational mechanisms reported earlier 3 . Most Peruvian haplotypes have been documented previously, except a seven-copy haplotype, which is within the expected realm of mutational possibilities but may harbor Peruvian-specific variants at the local scale. Discussion The amylase locus has long been discussed in the context of human dietary adaptation. Previous studies show that the salivary amylase gene ( AMY1) exhibits remarkable copy number variation, evolving rapidly and showing associations with dietary preferences among human populations 2 – 4 . However, direct evidence of local selection on AMY1 has remained elusive. Here, we provide robust evidence for recent selection acting on AMY1 copy number in Peruvian Andeans. Furthermore, we show that previously described recombination-based structural mutations, specifically those involving duplications or deletions of two AMY1 genes in a single mutational step, drive the formation of extremely high-copy AMY1 haplotypes in Peruvians. Our findings have broad implications for understanding the genetic basis of dietary adaptations to local ecologies and the role of selection in shaping structurally complex loci. Previous studies have shown that AMY1 copy number increased in the human lineage prior to agriculture 34 and has been shaped by frequent gains and losses. This same increase in amylase copy number has been demonstrated in species consuming a broad range diet that have cohabitated with humans after the agricultural revolution 5 – 7 . Thus, high-copy AMY1 haplotypes likely existed in populations ancestral to Andean Peruvians. SNV analyses on the flanking regions of the amylase locus indicate that the high-copy AMY1 haplotypes found in modern day Andean Peruvians are shared with high-copy haplotypes found in East Asians. Our results suggest that a soft selective sweep dramatically increased the allele frequency of these high-copy haplotypes in the Andean Peruvian population. This observation aligns with recent work on human adaptive evolution emphasizing the role of local selection on standing variation 35 , as demonstrated convincingly for skin color 36 , 37 . It also is noteworthy that AMY1 copy number variation emerges from recurrent NAHR events. The rates of these events are likely influenced by haplotype length where a higher number of AMY1 gene copies lead to a higher probability of NAHR. The selective sweep that led to higher allele frequencies of high-copy haplotypes may lead to additional de novo AMY1 copy number variation. This de novo variation could have been subject to selection alongside pre-existing variation. Therefore, it is plausible that both selection on standing variation and de novo mutations contributed to the elevated frequency of amylase haplotypes with five or more haploid AMY1 copies. This could have led to an increased proportion of Andean individuals with high-copy haplotype. An open question remains as to whether there is an upper limit to AMY1 copy number expansion and variation. Potential constraints may arise from genomic instability as a function of increased repeated segments or homologous alignment issues caused by copy number mismatches between maternal and paternal chromosomes. Investigating these limitations presents an exciting avenue for future research. Our observation that AMY1 copy number has adaptively increased in the Andean Peruvian population raises the broader question of the adaptive function of salivary amylase. At the molecular level, the enzyme’s primary role is starch digestion 38 , 39 , and there is accumulating evidence that increased AMY1 copy number leads to a higher dosage of functional paralogs, resulting in elevated salivary amylase levels 8 , 40 . However, how this increased enzyme dosage translates into organismal-level function remains unclear. Several studies have reported associations between higher AMY1 copy number and shifts in the oral and gut microbiomes. For instance, elevated copy number has been linked to greater oral microbiome diversity 8 including increased abundance of pathogenic Streptococcus species in the oral cavity, which may underlie the observed association between high AMY1 copy number and denture use in a large UK Biobank cohort 9 . In the gut, higher AMY1 dosage has been associated with enrichment of bacteria capable of digesting resistant starch 8 , potentially supporting more efficient metabolic homeostasis for starch-rich diets. Animal studies have shown that salivary amylase can trigger a cephalic-phase insulin response upon starch consumption 41 , potentially changing the energy metabolism under starch rich diets. In parallel, numerous studies have attempted to link AMY1 copy number variation to metabolic traits in humans, with inconsistent results. The largest and best-designed studies to date have not found a direct association with body mass index 18 , 42 , suggesting that any metabolic consequences are likely dependent on complex interactions involving diet, environment, and genetic background 43 . Within the cultural, dietary, and environmental context of the Andes, the evidence for selection on the amylase locus is concordant with previously suggested starch dietary adaptations among Andeans 16 , 17 . Potatoes were one of the first crops domesticated in the Andes, approximately 6,000-10,000 years ago 11 , 44 , 45 . Since then, their descendants have been consuming an extremely starch-rich; a cultural practice that continues today 13 , 46 . As mentioned above, researchers have found that the MGAM gene, which encodes the maltase-glucoamylase enzyme, shows evidence of selection in Andean populations 16 with our samples deriving from a subset of the individuals in Jorgensen et al. This enzyme functions downstream of amylase in the starch digestion pathway, breaking down oligosaccharides generated by amylase activity on complex carbohydrates. Thus, a compelling hypothesis is that the entire starch digestion pathway has been subject to positive selection in Indigenous Andean populations, shaped by a long history of extreme potato-starch intake. Increased AMY1 copy number, along with complementary changes in genes like MGAM , may have jointly optimized starch metabolism to support energy balance and life history traits. While our findings support adaptation to a starch-rich diet in Andean populations, this hypothesis carries important caveats. The reliance on starchy foods among Andean Peruvians is not vastly different from neighboring groups like the Maya, raising the question of what makes the Andean context unique. One possibility is that a higher AMY1 copy number may be beneficial for a potato-heavy diet, with potatoes among the highest in glycemic index among all human foods 12 . However, individuals with higher AMY1 copy number tend to have increased glycemic responses after consumption of potatoes 47 . Given that ∼60% of Andean Peruvians carry high AMY1 copy number haplotypes (≥ 10 copies), the potential for faster starch processing and regulation of glycemia may have conferred an advantage in this setting, but the precise organismal-level effects of this adaptation remain to be tested. We also observed similarly elevated AMY1 copy numbers in the Akimel O’odham of Southern Arizona and Northern Mexico. Although sample size precluded formal selection tests, this population presents a distinct ecological and dietary history, rooted in corn-based agriculture and desert environments 48 . If selection on starch metabolism occurred here, it likely reflects different pressures from the Andes, though many further studies are required. Despite their high AMY1 copy number, the Akimel O’odham, like the Andeans, suffer from high rates of metabolic disease today 49 , challenging the notion that this genetic variant protects against starch-driven metabolic disorders in modern diets. This pattern aligns with large-scale GWAS studies, which have failed to find strong associations between AMY1 copy number and obesity-related traits. Interestingly, the strongest biomedical correlate of high AMY1 copy number is poorer oral health 9 . This is consistent with the fact that in Peru, over 85% of school-aged children report cavities 50 . While these outcomes are influenced by social and behavioral factors, they highlight the complex health consequences of genetic adaptations shaped by past diets. By combining long-read sequencing with accurate CNV genotyping, our study offers a framework to study local adaptation at structurally complex loci. Similar approaches could illuminate other multigene families involved in digestion, such as PGA and LPA . The apparent selection on both AMY1 and MGAM in Andeans suggests potential co-adaptation across the starch digestion pathway. Future work examining how AMY1 and MGAM variants interact to process potato starch will help clarify their specific evolutionary roles. Parallel studies in other high-starch populations, ancient genomes, and functional assays will help test these hypotheses and broaden our understanding of the public health implications of AMY1 variation globally. Methods Sample collection Peruvian Andean study participants self-identified as Quechua and were recruited from Lima (14 m) and Cerro de Pasco (4,338 m), Peru ( Figure 1E ) as part of a larger study investigating high-altitude adaptation 21 , 51 . It is important to note that both Peruvian Quechua study participant groups were of high-altitude indigenous Andean ancestry. Study participants recruited in Lima, Peru, were born, raised, and resided at low altitudes, but whose parents and both sets of grandparents were of highland Quechua ancestry. Peruvian Quechua recruited in Cerro de Pasco, Peru, were born, raised, and resided at high altitudes. Indigenous Mexican Maya participants spoke Tzeltal, Tzotzil, or Ch’ol and were recruited around or in Palenque, Chiapas, Mexico (13 m). Male and female study participants were unrelated, healthy, non-pregnant/lactating, nonsmokers, and between the ages of 18 and 35. All study participants provided whole blood for DNA extraction. Study participants were recruited as part of a larger effort to broadly understand Andean evolutionary adaptation to high altitude. As such, Mexican Maya speakers were selected because 1. they are a close genetic relative of the Quechua who have never lived at high altitude and 2. Previous analysis has demonstrated that Indigenous Americans from Mexico, including Maya, are robust controls for identifying Andean signatures of selection 31 , 52 . Importantly for the study of selection at AMY1 , Andean and Mesoamerican populations diverged prior to the domestication of potatoes 53 . Furthermore, we were able to sample individuals from both populations (Quechua and Maya) with minimal evidence of recent European or African admixture. All participants provided written informed consent at enrollment for the study. The study was approved by the institutional review boards at the University of California Los Angeles (IRB # 20-001141), the University of Michigan (IRB #HUM00064387), Syracuse University (IRB#22-364), as well as The Research Ethics Committee of the Universidad Peruana Cayetano Heredia, Lima, Peru and the Centro de Investigación y Docencia Económicas, Mexico City, Mexico. Short-read genome FASTQ files from unrelated individuals in the 1000 Genomes Project (PRJEB31736) , the Human genome Diversity Project (HGDP) (PRJEB6463), along with samples from the Simons Genome Diversity Project (SGDP) (PRJEB9586) that belong to populations present in HGDP, were downloaded from the European Nucleotide Archive 27 , 28 , 54 . We additionally utilized high-coverage short-read sequence data from 38 Tibetan genomes 55 as well as high-coverage linked read data from 127 individuals from the Middle East 56 . For all of these datasets, we only kept populations with greater than or equal to eight individuals in order to maximize population diversity and minimize bias due to low sample size. Gene Copy Number Calling To calculate the amylase copy number, we employed two different read depth-based gene copy number callers: mrCaNaVaR and GeneToCN 57 – 59 . Both copy number callers have pros and cons. On the one hand, mrCaNaVaR is a promiscuous copy number caller that struggles to differentiate between closely related genes in the same gene families but allows for rapid genome-wide calculations. On the other hand, GeneToCN is quite adept at distinguishing between AMY1 and AMY2 genes as well as calculating accurate AMY1 copy numbers 3 , but requires much greater involvement in each different gene calculation. To this end, we have opted to use both in our analyses. For our mrCaNaVaR analyses, using the provided reference sequences and following the reported methods in 58 we first aligned each sample FASTQ file to the GRCh38 Human reference genome (CM000663.2) using mrsFAST. We then before calculated the copy number of all genes defined in the reference human_v38.nr.genes.bed file 60 . We have opted to use the GRCh38 reference genome instead of the newer T2T reference genome because GRCh38 includes three copies of AMY1 ( AMY1A , AMY1B , and AMY1C ), which is thought to represent both the ancestral version of the locus and the most common haplotype 3 , unlike the seven copies present in T2T. Unlike mrCaNaVaR and most other copy number callers, GeneToCN is an alignment-free method. Instead of calculating read depth across the entire genome, it focuses solely on the read depth of k-mers unique to the AMY1 genes and conserved flanking regions to determine gene copy number. Following the protocol described in 59 , we calculated the amylase copy number for AMY1 genes, AMY2B , and AMY2A using 25mers. VST outlier and median copy number analyses As the reference human_v38.nr.genes.bed file used in our mrCaNaVaR analysis contains other non-coding genetic features such as long non-coding RNAs, we filtered our dataset for all genes that overlapped with the Ensembl annotation database homo sapiens protein-coding genes from GRCh38 (version 97, ID “AH73881”). We then calculated the VST statistic, VST = (VT−VS)/VT ( see Github ), for all genes that had at least a copy number difference greater than 1 across the dataset in R version 4.4.1 using the protocol described in 19 to compare the genes between the populations PEL and MXL along with Quechua samples from Cerro de Pasco and Maya from Palenque. We did not include Quechua samples from Lima in this step due to the differing sequence coverage (∼15X coverage vs. ∼30X coverage). Following this step, we calculated the median of every gene in each population, subtracted the difference between the compared populations, and took the absolute value. After rounding the calculated gene depths and filtering for genes that had a median difference or VST value of 0, we produced a dataset of 576 genes for our PEL and MXL comparison and a dataset of 4,113 genes for our Quechua and Maya comparison. We hypothesized that the difference in coverage between the Quechua and Maya genomes and the PEL and MXL genomes (∼15X coverage vs. ∼30X coverage) might have contributed to a larger number of genes in the Quechua and Maya comparison as fewer reads might have induced greater variance in gene copy number for certain genes that are not in fact copy number variable. Admixture analyses We performed an unsupervised model in ADMIXTURE to determine genome-wide ancestry proportions using 826,848 SNVS after pruning for LD (r 2 > 0.1) and 1000 Genomes individuals as ancestral references 61 . Individuals were clustered with K=4 into the following super-populations of European, Asian, African, and Indigenous American ancestry. Locus-specific ancestry was assigned using RFMix2 and analyzed via chromosomal segments in windows corresponding to the AMY region 62 . The top 100 1KG individuals from each population with the most corresponding ancestry based on ADMIXTURE results were used as ancestral references for RFMix 2: European (IBS), Asian (CHB/CHS), African (YRI), and Indigenous American (MXL/PEL). Quechua & Maya VCF file creation Whole-genome data for the Quechua samples from Cerro de Pasco and Maya from Palenque, Chiapas, Mexico were generated using an Illumina NovaSeq with 75 base paired-end reads at 15× depth of coverage at the University of Michigan’s DNA Sequencing Core. Whole-genome data for Quechua samples from Lima were sequenced using DIPSEQ-T5 at a higher coverage (∼30X). Low-quality reads with a Phred score < 30 were removed following demultiplexing and trimming with Trim Galore! v.0.6.7 63 . Alignment was performed using BWA-MEM against hg38 64 , and variant calling was completed with GATK 65 . Reads were filtered and removed that had low mapping quality, missing per genotype > 0.10, or missing per individual > 0.10 using both VCFtools and Plink2.0 and then phased using SHAPEITv4.2 66 – 68 . FST and SNV/ AMY1 association We filtered out Quechua and Maya individuals with greater than 50% non-American ancestry for the amylase region based on the locus-specific ancestry analysis, removing three Quechua individuals from subsequent analyses (Quechua_Lowland_3, Quechua_Highland_23, & Quechua_Highland_24). Using chromosome 1 data, we filtered for missingness and indels using VCFtools 0.1.16, and retained only biallelic SNVs that had a minor allele frequency greater than 0.05. Following this step, we calculated FST between Quechua and Maya using VCFtools 69 . Information regarding the frequency of the alternative allele in each population was recorded and used to label if the alternative state was present more frequently in Maya or Quechua samples. For each SNV, we first applied a multiple-hypothesis corrected Kruskal–Wallis test 70 (s ee Github ) (Table S8 & S9 ). To control the family-wise error rate across all SNVs we adjusted the KW p-values with a Bonferroni correction 71 . Kruskal_pvalue <-vapply(snp_data, function(snv) { kruskal.test(AMY1_data ∼ snv)$p.value }, numeric(1)) We recorded information regarding the median AMY1 gene copy number for individuals containing homozygous reference, heterozygous, and homozygous alternative states for each SNV and calculated the median gene copy number for each group. Instances where the median AMY1 copy number of homozygous alternative individuals was greater than either the median of homozygous reference and/or heterozygous individuals, as well as instances where the median AMY1 copy number of heterozygous individuals was greater than homozygous reference individuals were multiplied by positive one. Instances where the median AMY1 copy number of homozygous reference individuals was greater than either the median of homozygous alternative and/or heterozygous individuals, as well as instances where the median AMY1 copy number of heterozygous individuals was greater than homozygous alternative individuals were multiplied by negative one to show a negative association with AMY1 copy number and the alternative SNV allele. We calculated a two-sided Wilcoxon test using “stat_compare_means” with comparisons between each SNV allele state (homozygous reference, heterozygous, and homozygous alternative) for two SNVs of interest 72 ( Figure 3B & 3D ). Though several of these SNVs are in and around the AMY2 genes, none of these SNVs are known to have functional effects on AMY1 or AMY2 expression or function and any intronic AMY2 SNVs that show high LD with selected AMY1 haplotypes are most likely hitchhiking passengers rather than targets of an independent selection event due to this close proximity. XP-EHH Analysis Using the same filtered SNV file from the FST and Kruskal-Wallis analyses, we calculated XP-EHH between the Quechua and Maya populations using selscan 2.0 73 . We created a genetic map file for these SNVs based on human chromosome 1 recombination rates recorded in 74 . These recombination rates are within the top 95th percentile of chromosome 1. To account for any associations between SNVs on either side of the amylase locus, we extended the haplotype extension to 2 million BP. We additionally applied the Weir and Goudet method in order to account for inbreeding, population structure, or genetic drift that may potentially inflate or distort selection signals 75 . Finally, we normalized our results across the chromosome. Rs143597860 ranked in the top 23/239,792 SNVs with a XP-EHH value of 6.4. Phylogeny We first filtered the phased SNV dataset for missingness, non-American ancestry at the amylase locus, and to only include biallelic SNVs using VCFtools 0.1.16 67 . We then created pseudo haploid FASTA files based on SNVs located between the regions of high recombination surrounding the amylase locus (chr1:103,347,982-103,830,994) for each individual using BCFtools 1.14, SAMtools 1.16.1, and GATK v4.3.0 65 , 76 . To do this, with each individual, we selected only variants where the individual has at least one copy of the alternative state. We then separated these SNVs based on the haploid genome and used the GRCh38 reference genome to create a haploid FASTA file. gatk SelectVariants \ -R chr1_hg38.fasta \ -V${sample}.vcf.gz \ --select-type-to-include SNP \ --exclude-non-variants true \ --O ${sample}_excluded_nonvariants.vcf.gz # For haplotype 1 samtools faidx chr1_hg38.fasta ${location} | bcftools consensus -H 1 ${sample}_excluded_nonvariants.vcf.gz > ${sample}_region_hap1.fasta After merging all of the FASTA files, we constructed a maximum likelihood tree by testing multiple models, applying robust branch supports (ultrafast bootstrap), and performing an exhaustive NNI tree search using IQ-TREE 2.2.6 77 – 79 . The output .treefile was labeled with the populations and AMY1 copy numbers in iTOL version 6 80 . iqtree -s LWK_Quechua_Maya_no_missing_biallelic.fasta --seqtype DNA -m TESTMERGE -bb 1000 -bnni -allnni -nt AUTO -pre Quechua_Maya_no_missing_biallelic_nni Since the phylogeny was constructed using haploid, phased SNV data mapped to the hg38 reference genome, each tip in the tree represents a single haploid genome per individual. However, AMY1 copy number is measured on a diploid basis. Therefore, individuals with high AMY1 copy numbers who appear outside the high-copy-number clade in the phylogeny may carry one high-copy haplotype (from the clade) and one low-copy haplotype. This apparent mismatch arises because the phylogeny captures only one haplotype per individual, while the copy number data reflects both. As a result, some individuals with high total copy numbers may possess one high-copy haplotype from the identified clade and another low-copy haplotype present elsewhere in the phylogeny. This interpretation aligns with the data, as 13 out of 14 Quechua SNV haplotypes associated with a diploid AMY1 copy number of 14, an unusually high count requiring an average of seven copies per haplotype, fall within this subclade. Building pangenome We constructed a local pangenome graph to investigate the mutational mechanisms underlying structural variation at the amylase locus with three specific aims. First, although non-allelic homologous recombination (NAHR) is the dominant mechanism driving AMY1 copy number variation, other mechanisms such as microhomology-mediated break-induced replication (MMBIR) have been shown to produce structural haplotypes at this locus, including even-copy AMY1 alleles and AMY2 variants 3 . We sought to test whether MMBIR contributed to structural diversity in Peruvian populations. Second, given the recurrent nature of AMY1 duplications, we aimed to confirm whether previously described recombination breakpoints also underlie the Peruvian haplotypes. Third, we used this graph-based approach to identify potential uncommon haplotypes, including the 7-copy AMY1 haplotype we report here. The pangenome was built using 40 known haplotypes comprising the entire AMY locus 2 , 3 . For haplotypes HG02004_pat_H9, HG02293_mat_H5, HG02293_mat_H5, HG02004_mat_H3, HG01978_mat_H3, HG02300_mat_H3, HG02148_pat_H3, HG02293_pat_H1, HG02300_pat_H3, HG01106_mat_H7 2 regions comprising the amy locus were extracted by blasting regions upstream and downstream of the locus. Human reference genome hg38 was used to retrieve the upstream and downstream sequences, respectively chr1:103,547,648-103,560,699 and chr1:103,766,322-103,785,201. For haplotypes H1.1, H1.2, H2A0.1, H2A2B2.1, H3r.1, H3r.2, H3r.3, H3r.4, H3r.5, H3r.6, H3r.7, H3B2.1, H3A2.1, H3A3B3.1, H4A0.1, H4A0.2, H4A2.1, H4A2.2, H4A2B2.1, H5.1, H5.2, H5.3, H5.4, H5.5, H5.6, H5.7, H5.8, H5.9, H5.10, and H7.1, data was retrieved from 3 ) (Supplementary Table 8). All data and scripts used for the pangenome can be found in Github. We used PGGB 81 to build a local pangenome with the following parameters: pggb -i amylocus_haplotypes.fa -o amylocus_5k -n 46 -t 32 -p 99 -s 5k Long-read sequencing and assembly To understand the Peruvian haplotypes for the amylase locus, we chose six Peruvian samples from the 1000 Genomes Project 28 to sequence (HG01975, HG01972, HG02106, HG0946, HG01922, HG01955). The samples were chosen to represent the diversity of the population, so we chose samples representing 6, 8, 14, and 18 diploid AMY1 copies, respectively. We generated long-reads on the Oxford Nanopore PromethION sequencing platform using the SQK-LSK114 chemistry and R10.4.1 flow cells for all samples. Because this locus was complicated to assemble correctly, we also generated ultra-long reads for samples HG01975 and HG01946. We corrected the reads using HERRO 82 and assembled the genomes using hifiasm 83 . We adapted the hifiasm pipeline for the use of corrected ONT reads, as described in 82 . seqkit sliding -s 30000 -W 30000 -g ${cor_reads} > chopped.fasta seqkit seq -m 10000 chopped.fasta > processed.fasta #As ultra-long reads, we are using only reads longer than 50kb seqkit seq -m 50000 $raw > raw_reads_${sample}_50kb.fastq ./hifiasm -t 32 -o ${sample}_diploid_50kb --ul raw_reads_${sample}_50kb.fastq -- ul-cut 50000 --dual-scaf -1 parent1.yak -2 parent2.yak processed.fasta To verify the assembly recovery of the amylase locus, we aligned the assemblies to the publicly available genome hg38 (Figure S15 ). Based on the dot plots and manual inspection, we concluded that the assembled haplotypes were not correctly assembled. We then used the contigs that span the region to recover the phased reads using the following approach: 1) we mapped all the contigs to the hg38 reference genome; 2) we retrieved the contigs that mapped to the AMY locus (coordinates); 3) we retrieved the reads that mapped to the specific parts of the contigs that mapped to the region of interest. All scripts used for this are on Github . Mapping to haplotypes and filtering Because the assembled contigs were incorrect, we mapped the reads to all known distinct haplotypes (n=30) plus previously published Peruvian haplotypes 2 and used a read depth measure to decide the best fit for each sample. The haplotypes used are on Github , details in Table S15 . We competitively mapped the reads into each haplotype and filtered the mapping for length > 2.5kb. This filtering was done because the region is rich in transposable elements that might skew the analysis. We hypothesized that the best-fit haplotype will have a lower variance in mapping. Therefore, we calculated the mean and standard deviation for the read depth on the entire haplotype. The best fit for all haplotypes was the one that showed the lowest standard deviation and a mean read depth of ≥ 1. We also checked that the best-fit haplotype for each individual coincided with the estimated copy number based on ddPCR results ( Table S13 ). Based on this method, we could assign the best fit for 4 individuals (8 phased haplotypes). However, two individuals (HG02106 and HG01975) were assigned haplotypes discordant with the estimated copy number generated by ddPCR. We manually inspected the best-fit haplotypes, and the mapping was not homogeneous for any of the top haplotypes (H9, H7, H5). The best match for this haplotype was H9, and, therefore, we hypothesized that the best match would be a haplotype similar to H9 but harboring 7 AMY1 copies. We generated a mock H7 haplotype by deleting one duplication unit ( AMY1A and AMY1B , see 3 ) from the H9 haplotype ( Supplementary Methods, Figure S16 ). Data availability Results generated for this study are provided as Supplementary Data 1-15. Additional datasets are provided in this paper including the PBS results, the Ohana results, the XP-EHH results, the mock H7 FASTA file, and the SNV LD analyses have been uploaded to Github https://github.com/kendrasc/Andean_Amylase_Project . Phased SNV data between the locations of Chr1: 103343000-103960000 in the Quechua and Maya individuals is available on Figshare along with archived versions of the aforementioned results and datasets where they are freely available without restriction. The 1KG data and the HGDP data can be obtained at https://www.internationalgenome.org/data/ . The SGDP data can be obtained at https://www.simonsfoundation.org/simons-genome-diversity-project/ . The raw sequencing data generated for this study have been deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession number PRJNA1240630. The FASTQ files are available under the following SRA accession numbers: HG02106 (SRR32828259), HG01946 (SRR32828263), HG02252 (SRR32828258), HG01975 (SRR32828261), HG01972 (SRR32828257), HG01922 (SRR32828264), HG02006 (SRR32828260), HG01955 (SRR32828262). Each sample is associated with a BioSample record (accessions SAMN47540692 to SAMN47540699) and represents Homo sapiens individuals of Peruvian ancestry from the 1000 Genomes Project. For peer review, the data can be accessed via the following private link: https://dataview.ncbi.nlm.nih.gov/object/PRJNA1240630?reviewer=2reo2tk31pprot2497moukd0e2 The whole genome data that support the findings of this study may be available upon reasonable request to the corresponding authors. Code availability The code used to generate the results in this study is publicly accessible on GitHub https://github.com/kendrasc/Andean_Amylase_Project . Acknowledgments We are incredibly grateful to the study participants from Lima, Peru, Cerro de Pasco, Peru, and Palenque, Chiapas, Mexico. We are also thankful to Dr. Maria Inclan (Centro de Investigación y Docencia Económicas), Dr Sudipta Ghosh (NEHU University, Shillong, India), Nate Bartman, Jason Howard, Jacqueline Imse, Kevin Heffernan (Syracuse University), Mark Olfert (University of West Virginia), Francisco Villafuerte (Universidad Peruana Cayetano Heredia), Katarina Evans (CUNY), Neha Angel, Thea Spindel, Sarah Burke, Sarita Greer, Adam Conner, Isabel Meda, and Alber Aqil. This work was supported by NSF grants BIO-2049947 AND BCS-2123284 (to O.G.), BCS-1132310 (to A.W.B., T.D.B., and F.L.-V.) and BCS-1638642 (to A.W.B., T.D.B., and F.S.L.), NIH-R35GM156519 (to O.G.), Leakey Foundation Grant (to A.W.B.), and an Elizabeth Caroline Crosby Grant from the University of Michigan (to A.W.B.). Funder Information Declared National Science Foundation, https://ror.org/021nxhr62 , BIO-2049947 , BCS-2123284 , BCS-1132310 , BCS-1638642 Leakey Foundation, https://ror.org/018kdpd27 Footnotes In this version, Figures 1-4 have been update. There have also added several additional analyses. https://github.com/kendrasc/Andean_Amylase_Project References 1. ↵ Roberts , P. J. P. & Whelan , W. J . The mechanism of carbohydrase action. 5. Action of human salivary α-amylase on amylopectin and glycogen* . Biochem. J. 76 , 246 – 253 ( 1960 ). OpenUrl FREE Full Text 2. ↵ Bolognini , D. et al. Recurrent evolution and selection shape structural diversity at the amylase locus . Nature 634 , 617 – 625 ( 2024 ). OpenUrl CrossRef PubMed 3. ↵ Yilmaz , F. et al. Reconstruction of the human amylase locus reveals ancient duplications seeding modern-day variation . Science 386 , ( 2024 ). 4. ↵ Perry , G. H. et al. Diet and the evolution of human amylase gene copy number variation . Nat. Genet . 39 , 1256 – 1260 ( 2007 ). OpenUrl CrossRef PubMed Web of Science 5. ↵ Pajic , P. et al. Independent amylase gene copy number bursts correlate with dietary preferences in mammals . eLife 8 , e44628 ( 2019 ). OpenUrl CrossRef PubMed 6. Arendt , M. , Fall , T. , Lindblad-Toh , K. & Axelsson , E . Amylase activity is associated with 2B copy numbers in dog: implications for dog domestication, diet and diabetes . Anim. Genet . 45 , 716 – 722 ( 2014 ). OpenUrl CrossRef PubMed 7. ↵ Axelsson , E. et al. The genomic signature of dog domestication reveals adaptation to a starch-rich diet . Nature 495 , 360 – 364 ( 2013 ). OpenUrl CrossRef PubMed Web of Science 8. ↵ Poole , A. C. et al. Human Salivary Amylase Gene Copy Number Impacts Oral and Gut Microbiomes . Cell Host Microbe 25 , 553 – 564 .e7 ( 2019 ). OpenUrl CrossRef PubMed 9. ↵ Kamitaki , N. et al. Human and bacterial genetic variation shape oral microbiomes and health . Preprint at doi: 10.1101/2025.03.31.25324952 ( 2025 ). OpenUrl Abstract / FREE Full Text 10. ↵ Haas Jr , W. R. H. & Llave , C. V. Hunter-gatherers on the eve of agriculture: investigations at Soro Mik’aya Patjxa, Lake Titicaca Basin, Peru, 8000–6700 BP . Antiquity 89 , 1297 – 1312 ( 2015 ). OpenUrl CrossRef 11. ↵ Spooner , D. M. , McLean , K. , Ramsay , G. , Waugh , R. & Bryan , G. J . A single domestication for potato based on multilocus amplified fragment length polymorphism genotyping . Proc. Natl. Acad. Sci . 102 , 14694 – 14699 ( 2005 ). OpenUrl Abstract / FREE Full Text 12. ↵ Atkinson , F. S. , Brand-Miller , J. C. , Foster-Powell , K. , Buyken , A. E. & Goletzke , J . International tables of glycemic index and glycemic load values 2021: a systematic review . Am. J. Clin. Nutr . 114 , 1625 – 1632 ( 2021 ). OpenUrl CrossRef PubMed 13. ↵ Berti , P. R. et al. Assessment and characterization of the diet of an isolated population in the Bolivian Andes . Am. J. Hum. Biol . 22 , 741 – 749 ( 2010 ). OpenUrl CrossRef PubMed 14. Chen , J. C. et al. Stable isotope chemistry reveals plant-dominant diet among early foragers on the Andean Altiplano, 9.0–6.5 cal. ka . PLOS ONE 19 , e0296420 ( 2024 ). OpenUrl CrossRef PubMed 15. ↵ Rose , D. , Burgos , G. , Bonierbale , M. & Thiele , G . Understanding the role of potatoes in the Peruvian diet: An approach that combines food composition with household expenditure data . J. Food Compos. Anal . 22 , 525 – 532 ( 2009 ). OpenUrl 16. ↵ Jorgensen , K. , Garcia , O. A. , Kiyamu , M. , Brutsaert , T. D. & Bigham , A. W . Genetic adaptations to potato starch digestion in the Peruvian Andes . Am. J. Biol. Anthropol . 180 , 162 – 172 ( 2023 ). OpenUrl PubMed 17. ↵ Lindo , J. et al. The genetic prehistory of the Andean highlands 7000 years BP though European contact . Sci. Adv . 4 , eaau4921 ( 2018 ). OpenUrl FREE Full Text 18. ↵ Usher , C. L. et al. Structural forms of the human amylase locus and their relationships to SNPs, haplotypes and obesity . Nat. Genet . 47 , 921 – 925 ( 2015 ). OpenUrl CrossRef PubMed 19. ↵ Redon , R. et al. Global variation in copy number in the human genome . Nature 444 , 444 – 454 ( 2006 ). OpenUrl CrossRef PubMed Web of Science 20. ↵ Sabeti , P. C. et al. Genome-wide detection and characterization of positive selection in human populations . Nature 449 , 913 – 918 ( 2007 ). OpenUrl CrossRef PubMed Web of Science 21. ↵ Brutsaert , T. D. et al. Association of EGLN1 gene with high aerobic capacity of Peruvian Quechua at high altitude . Proc. Natl. Acad. Sci . 116 , 24006 – 24011 ( 2019 ). OpenUrl Abstract / FREE Full Text 22. ↵ Yi , X. et al. Sequencing of 50 human exomes reveals adaptation to high altitude . Science 329 , 75 – 78 ( 2010 ). OpenUrl Abstract / FREE Full Text 23. ↵ Cheng , J. Y. , Mailund , T. & Nielsen , R . Fast admixture analysis and population tree estimation for SNP and NGS data . Bioinformatics 33 , 2148 – 2155 ( 2017 ). OpenUrl CrossRef PubMed 24. ↵ Raghavan , M. et al. Genomic evidence for the Pleistocene and recent population history of Native Americans . Science 349 , aab3884 ( 2015 ). OpenUrl Abstract / FREE Full Text 25. ↵ Willerslev , E. & Meltzer , D. J . Peopling of the Americas as inferred from ancient genomics . Nature 594 , 356 – 364 ( 2021 ). OpenUrl CrossRef PubMed 26. ↵ Albers , P. K. & McVean , G . Dating genomic variants and shared ancestry in population-scale sequencing data . PLOS Biol . 18 , e3000586 ( 2020 ). OpenUrl CrossRef PubMed 27. ↵ Mallick , S. et al. The Simons Genome Diversity Project: 300 genomes from 142 diverse populations . Nature 538 , 201 – 206 ( 2016 ). OpenUrl CrossRef PubMed 28. ↵ The 1000 Genomes Project Consortium et al. A global reference for human genetic variation . Nature 526 , 68 – 74 ( 2015 ). OpenUrl CrossRef PubMed 29. ↵ Speidel , L. , Forest , M. , Shi , S. & Myers , S. R . A method for genome-wide genealogy estimation for thousands of samples . Nat. Genet . 51 , 1321 – 1329 ( 2019 ). OpenUrl CrossRef PubMed 30. ↵ Stern , A. J. , Wilton , P. R. & Nielsen , R . An approximate full-likelihood method for inferring selection and allele frequency trajectories from DNA sequence data . PLOSGenet . 15 , e1008384 ( 2019 ). OpenUrl 31. ↵ Bigham , A. W. et al. Identifying positive selection candidate loci for high-altitude adaptation in Andean populations . Hum. Genomics 4 , ( 2009 ). 32. ↵ Fehren-Schmitz , L. & Georges , L . Ancient DNA reveals selection acting on genes associated with hypoxia response in pre-Columbian Peruvian Highlanders in the last 8500 years . Sci. Rep . 6 , ( 2016 ). 33. ↵ Logsdon , G. A. et al. Complex genetic variation in nearly complete human genomes . Nature 1 – 12 ( 2025 ) doi: 10.1038/s41586-025-09140-6 . OpenUrl CrossRef 34. ↵ Inchley , C. E. et al. Selective sweep on human amylase genes postdates the split with Neanderthals . Sci. Rep . 6 , 37198 ( 2016 ). 35. ↵ Schrider , D. R. & Kern , A. D . Soft Sweeps Are the Dominant Mode of Adaptation in the Human Genome . Mol. Biol. Evol . 34 , 1863 – 1877 ( 2017 ). OpenUrl CrossRef PubMed 36. ↵ Crawford , N. G. et al. Loci associated with skin pigmentation identified in African populations . Science 358 , eaan8433 ( 2017 ). OpenUrl Abstract / FREE Full Text 37. ↵ Martin , A. R. et al. An Unexpectedly Complex Architecture for Skin Pigmentation in Africans . Cell 171 , 1340 – 1353 .e14 ( 2017 ). OpenUrl CrossRef PubMed 38. ↵ Butterworth , P. J. , Warren , F. J. & Ellis , P. R . Human α-amylase and starch digestion: An interesting marriage . Starch - Stärke 63 , 395 – 405 ( 2011 ). OpenUrl 39. ↵ Van Der Maarel , M. J. E. C. , Van Der Veen , B. , Uitdehaag , J. C. M. , Leemhuis , H. & Dijkhuizen , L . Properties and applications of starch-converting enzymes of the α-amylase family . J. Biotechnol . 94 , 137 – 155 ( 2002 ). OpenUrl CrossRef PubMed Web of Science 40. ↵ Mandel , A. L. , Gachons, C. P. des, Plank , K. L. , Alarcon , S. & Breslin , P. A. S. Individual Differences in AMY1 Gene Copy Number, Salivary α-Amylase Levels, and the Perception of Oral Starch . PLOS ONE 5 , e13352 ( 2010 ). OpenUrl CrossRef PubMed 41. ↵ Glendinning , J. I. , Lubitz , G. S. & Shelling , S . Taste of glucose elicits cephalic-phase insulin release in mice . Physiol. Behav . 192 , 200 – 205 ( 2018 ). OpenUrl PubMed 42. ↵ Shwan , N. A. A. & Armour , J. A. L . No Evidence for Association of BMI with Salivary Amylase Gene Copy Number in the UK 1958 Birth Cohort . Obesity 27 , 1533 – 1538 ( 2019 ). OpenUrl PubMed 43. ↵ Rossi , N. et al. Ethnic-specific association of amylase gene copy number with adiposity traits in a large Middle Eastern biobank . Npj Genomic Med . 6 , ( 2021 ). 44. ↵ Li , Y. et al. Genomic Analyses Yield Markers for Identifying Agronomically Important Genes in Potato . Mol. Plant 11 , 473 – 484 ( 2018 ). OpenUrl CrossRef PubMed 45. ↵ Rumold , C. U. & Aldenderfer , M. S . Late Archaic–Early Formative period microbotanical evidence for potato at Jiskairumoko in the Titicaca Basin of southern Peru . Proc. Natl. Acad. Sci . 113 , 13672 – 13677 ( 2016 ). OpenUrl Abstract / FREE Full Text 46. ↵ De Haan , S. et al. Traditional Processing of Black and White Chuño in the Peruvian Andes: Regional Variants and Effect on the Mineral Content of Native Potato Cultivars . Econ. Bot . 64 , 217 – 234 ( 2010 ). OpenUrl 47. ↵ Atkinson , F. S. , Hancock , D. , Petocz , P. & Brand-Miller , J. C . The physiologic and phenotypic significance of variation in human amylase gene copy number . Am. J. Clin.Nutr . 108 , 737 – 748 ( 2018 ). OpenUrl CrossRef PubMed 48. ↵ Hunt , R. C. & Ingram , S. E . FOOD PRODUCTION CALENDAR FOR THE MIDDLE GILA RIVER, ARIZONA: AKIMEL O’ODHAM (PIMA) AND HOHOKAM . KIVA 79 , 253 – 279 ( 2014 ). OpenUrl 49. ↵ Schulz , L. O. & Chaudhari , L. S . High-Risk Populations: The Pimas of Arizona and Mexico . Curr. Obes. Rep . 4 , 92 – 98 ( 2015 ). OpenUrl PubMed 50. ↵ Castillo , J. L. , Palma , C. & Cabrera-Matta , A . Early Childhood Caries in Peru . Front. Public Health 7 , ( 2019 ). 51. ↵ Childebayeva , A. et al. Genome-Wide Epigenetic Signatures of Adaptive Developmental Plasticity in the Andes . Genome Biol. Evol . 13 , evaa239 ( 2021 ). OpenUrl CrossRef PubMed 52. ↵ Bigham , A. et al. Identifying Signatures of Natural Selection in Tibetan and Andean Populations Using Dense Genome Scan Data . PLoS Genet . 6 , e1001116 ( 2010 ). OpenUrl CrossRef PubMed 53. ↵ Kennett , D. J. et al. South-to-north migration preceded the advent of intensive farming in the Maya region . Nat. Commun . 13 , 1530 ( 2022 ). OpenUrl PubMed 54. ↵ Bergström , A. et al. Insights into human genetic variation and population history from 929 diverse genomes . Science 367 , eaay5012 ( 2020 ). OpenUrl Abstract / FREE Full Text 55. ↵ Lu , D. et al. Ancestral Origins and Genetic History of Tibetan Highlanders . Am. J. Hum. Genet . 99 , 580 – 594 ( 2016 ). OpenUrl CrossRef PubMed 56. ↵ Almarri , M. A. et al. The genomic history of the Middle East . Cell 184 , 4612 – 4625 .e14 ( 2021 ). OpenUrl CrossRef PubMed 57. ↵ Alkan , C. et al. Personalized copy number and segmental duplication maps using next-generation sequencing . Nat. Genet . 41 , 1061 – 1067 ( 2009 ). OpenUrl CrossRef PubMed Web of Science 58. ↵ Kahveci , F. & Alkan , C . Whole-Genome Shotgun Sequence CNV Detection Using Read Depth . in Copy Number Variants: Methods and Protocols (ed. Bickhart , D. M. ) 61 – 72 ( Springer, New York, NY , 2018 ). doi: 10.1007/978-1-4939-8666-8_4 . OpenUrl CrossRef 59. ↵ Pajuste , F.-D. & Remm , M . GeneToCN: an alignment-free method for gene copy number estimation directly from next-generation sequencing reads . Sci. Rep . 13 , 17765 ( 2023 ). 60. ↵ Hach , F. et al. mrsFAST: a cache-oblivious algorithm for short-read mapping . Nat.Methods 7 , 576 – 577 ( 2010 ). OpenUrl 61. ↵ Alexander , D. H. & Lange , K . Enhancements to the ADMIXTURE algorithm for individual ancestry estimation . BMC Bioinformatics 12 , 246 ( 2011 ). 62. ↵ Maples , B. K. , Gravel , S. , Kenny , E. E. & Bustamante , C. D . RFMix: A Discriminative Modeling Approach for Rapid and Robust Local-Ancestry Inference . Am. J. Hum. Genet . 93 , 278 – 288 ( 2013 ). OpenUrl CrossRef PubMed 63. ↵ Martin , M . Cutadapt removes adapter sequences from high-throughput sequencing reads . EMBnet.journal 17 , 10 – 12 ( 2011 ). OpenUrl 64. ↵ Li , H. & Durbin , R . Fast and accurate short read alignment with Burrows–Wheeler transform . Bioinformatics 25 , 1754 – 1760 ( 2009 ). OpenUrl CrossRef PubMed Web of Science 65. ↵ DePristo , M. A. et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data . Nat. Genet . 43 , 491 – 498 ( 2011 ). OpenUrl CrossRef PubMed Web of Science 66. ↵ Chang , C. C. et al. Second-generation PLINK: rising to the challenge of larger and richer datasets . GigaScience 4 , s13742-015-0047–8 ( 2015 ). 67. ↵ Danecek , P. et al. The variant call format and VCFtools . Bioinformatics 27 , 2156 – 2158 ( 2011 ). OpenUrl CrossRef PubMed Web of Science 68. ↵ Delaneau , O. , Zagury , J.-F. , Robinson , M. R. , Marchini , J. L. & Dermitzakis , E. T . Accurate, scalable and integrative haplotype estimation . Nat. Commun . 10 , 5436 ( 2019 ). OpenUrl CrossRef PubMed 69. ↵ Weir , B. S. & Cockerham , C. C . ESTIMATING F-STATISTICS FOR THE ANALYSIS OF POPULATION STRUCTURE . Evol. Int. J. Org. Evol . 38 , 1358 – 1370 ( 1984 ). OpenUrl 70. ↵ Kruskal , W. H. & Wallis , W. A . Use of Ranks in One-Criterion Variance Analysis . J. Am. Stat. Assoc . 47 , 583 – 621 ( 1952 ). OpenUrl CrossRef Web of Science 71. ↵ Carlo E. Bonferroni . Teoria statistica delle classi e calcolo delle probabilità . 62 ( 1936 ). 72. ↵ Wilcoxon , F . Individual Comparisons by Ranking Methods . Biom. Bull . 1 , 80 – 83 ( 1945 ). OpenUrl CrossRef 73. ↵ Szpiech , Z . A. selscan 2.0: scanning for sweeps in unphased data . Bioinformatics 40 , btae006 ( 2024 ). OpenUrl CrossRef PubMed 74. ↵ Halldorsson , B. V. et al. Characterizing mutagenic effects of recombination through a sequence-level genetic map . Science 363 , eaau1043 ( 2019 ). OpenUrl Abstract / FREE Full Text 75. ↵ Weir , B. S. & Goudet , J . A Unified Characterization of Population Structure and Relatedness . Genetics 206 , 2085 – 2103 ( 2017 ). OpenUrl Abstract / FREE Full Text 76. ↵ Danecek , P. et al. Twelve years of SAMtools and BCFtools . GigaScience 10 , giab008 ( 2021 ). OpenUrl CrossRef PubMed 77. ↵ Hoang , D. T. , Chernomor , O. , von Haeseler , A. , Minh , B. Q. & Vinh , L. S . UFBoot2: Improving the Ultrafast Bootstrap Approximation . Mol. Biol. Evol . 35 , 518 – 522 ( 2018 ). OpenUrl CrossRef PubMed 78. Kalyaanamoorthy , S. , Minh , B. Q. , Wong , T. K. F. , von Haeseler , A. & Jermiin , L. S . ModelFinder: fast model selection for accurate phylogenetic estimates . Nat. Methods 14 , 587 – 589 ( 2017 ). OpenUrl CrossRef PubMed 79. ↵ Minh , B. Q. et al. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era . Mol. Biol. Evol . 37 , 1530 – 1534 ( 2020 ). OpenUrl CrossRef PubMed 80. ↵ Letunic , I. & Bork , P . Interactive Tree of Life (iTOL) v6: recent updates to the phylogenetic tree display and annotation tool . Nucleic Acids Res . 52 , W78 – W82 ( 2024 ). OpenUrl CrossRef PubMed 81. ↵ Garrison , E. et al. Building pangenome graphs . Nat. Methods 21 , 2008 – 2012 ( 2024 ). OpenUrl CrossRef PubMed 82. ↵ Stanojević , D. , Lin , D. , Nurk , S. , Sessions, P. F. de & Šikić , M. Telomere-to-Telomere Phased Genome Assembly Using HERRO-Corrected Simplex Nanopore Reads . 2024.05.18.594796 Preprint at doi: 10.1101/2024.05.18.594796 ( 2024 ). OpenUrl Abstract / FREE Full Text 83. ↵ Cheng , H. , Concepcion , G. T. , Feng , X. , Zhang , H. & Li , H . Haplotype-resolved de novo assembly using phased assembly graphs with hifiasm . Nat. Methods 18 , 170 – 175 ( 2021 ). OpenUrl CrossRef PubMed View the discussion thread. Back to top Previous Next Posted August 15, 2025. Download PDF Supplementary Material Data/Code Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following Rapid Adaptive Increase of Amylase Gene Copy Number in Indigenous Andeans Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share Rapid Adaptive Increase of Amylase Gene Copy Number in Indigenous Andeans Kendra Scheer , Luane J. B. Landau , Kelsey Jorgensen , Charikleia Karageorgiou , Lindsey Siao , Can Alkan , Angelis M. Morales-Rivera , Christopher Osborne , Obed Garcia , Laurel Pearson , Melisa Kiyamu , María Rivera-Chira , Fabiola Leon-Velarde , Frank Lee , Tom Brutsaert , Abigail W. Bigham , Omer Gokcumen bioRxiv 2025.03.25.644684; doi: https://doi.org/10.1101/2025.03.25.644684 Share This Article: Copy Citation Tools Rapid Adaptive Increase of Amylase Gene Copy Number in Indigenous Andeans Kendra Scheer , Luane J. B. Landau , Kelsey Jorgensen , Charikleia Karageorgiou , Lindsey Siao , Can Alkan , Angelis M. Morales-Rivera , Christopher Osborne , Obed Garcia , Laurel Pearson , Melisa Kiyamu , María Rivera-Chira , Fabiola Leon-Velarde , Frank Lee , Tom Brutsaert , Abigail W. Bigham , Omer Gokcumen bioRxiv 2025.03.25.644684; doi: https://doi.org/10.1101/2025.03.25.644684 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Genetics Subject Areas All Articles Animal Behavior and Cognition (7618) Biochemistry (17635) Bioengineering (13859) Bioinformatics (41846) Biophysics (21401) Cancer Biology (18534) Cell Biology (25422) Clinical Trials (138) Developmental Biology (13352) Ecology (19860) Epidemiology (2067) Evolutionary Biology (24285) Genetics (15582) Genomics (22463) Immunology (17700) Microbiology (40298) Molecular Biology (17141) Neuroscience (88424) Paleontology (666) Pathology (2825) Pharmacology and Toxicology (4813) Physiology (7633) Plant Biology (15107) Scientific Communication and Education (2042) Synthetic Biology (4284) Systems Biology (9808) Zoology (2267)