https://orcid.org/0000-0003-0278-7655
Figures
Abstract
The association between salivary amylase gene (AMY1) copy number (CN) and metabolic health parameters, such as insulin resistance, has been studied but with conflicting results. The etiology of these inconsistencies is multifaceted and confounded by differences in genotyping methods. In this study, we investigated the factors that affect the association between AMY1 CN, salivary amylase activity (SAA), and type 2 diabetes (T2D) status. We collected up to four saliva samples from two cohorts: individuals with self-reported T2D or prediabetes (n = 18) and individuals who self-reported as being without T2D or prediabetes (controls, n = 178). We genotyped individuals for AMY1 CN using both quantitative PCR and droplet digital PCR and measured SAA in the saliva samples. We demonstrated that these two commonly used methods give comparable values for AMY1 CN. We showed a positive association between the time of saliva collection and SAA throughout the day. We also observed that AMY1 CN and SAA are positively associated in individuals with and without T2D or prediabetes. With increasing AMY1 CN, SAA was higher for each additional copy of AMY1 CN in individuals with T2D or prediabetes compared to controls. Our findings support the premise that increased SAA in T2D patients is a compensatory mechanism related to the long-term protective effect of high AMY1 CN on glucose metabolism.
Citation: Devarakonda SLS, Ren J, Poole AC (2025) The association between salivary amylase gene copy number and enzyme activity with type 2 diabetes status. PLoS One 20(7): e0324660. https://doi.org/10.1371/journal.pone.0324660
Editor: Gyaneshwer Chaubey, Banaras Hindu University, INDIA
Received: December 27, 2024; Accepted: April 29, 2025; Published: July 2, 2025
Copyright: © 2025 Devarakonda et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data is available on Zenodo: https://doi.org/10.5281/zenodo.15054014
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Copy number (CN) variation affects ~10% of the human genome [1]. CN variants (CNVs) modulate gene expression [2,3] and, as such, are genetic causes of phenotypic variability and disease risk. However, reliable determination of gene CN is challenging. One well known CNV is AMY1, which encodes the salivary amylase enzyme. Amylase is the most abundant enzyme in human saliva and initiates starch degradation in the mouth. In humans, diploid AMY1 CN ranges from 2 to 20 [4], and this variation is thought to affect metabolic health [5–8]. However, there is conflicting evidence regarding the relationship between AMY1 CN and metabolic disorders such as obesity and impaired glucose homeostasis [9–13]. These inconsistencies in the associations between AMY1 CN and metabolic health could be partly attributed to the methodological differences in determining AMY1 CN.
Studies involving AMY1 have used various methods to determine CN, including quantitative PCR (qPCR) [8,11,14,15], droplet digital PCR (ddPCR) [5,6,10], fluorescence in situ hybridization (FISH) [9,14], paralogue ratio tests (PRT) [9,16,17], and computational approaches [10]. Research groups using these methods to address the same question have come to different conclusions. For example, using qPCR, one of the first studies addressing the association between AMY1 CN and BMI reported an inverse relationship [11]. By contrast, because of the correlation between AMY1 CN and historic dietary patterns, an earlier publication proposed that a higher CN conferred the fitness advantage of weight gain due to more efficient extraction of calories from starchy foods [14]—this supports the premise that AMY1 CN is positively correlated with BMI. Other research groups, using methods other than qPCR to determine AMY1 CN, reported no association between AMY1 CN and BMI [9,10]. This provoked questions regarding the accuracy of qPCR versus other methods for AMY1 CN determination [18]. The first goal of our study was to compare qPCR and ddPCR, two commonly used laboratory-based methods to determine AMY1 CN; ddPCR is known for accuracy but qPCR is still sometimes used depending on resources available to research teams.
It has been reported that AMY1 CN is positively correlated with salivary amylase protein quantity and SAA [19]. However, assuming a consistent correlation between AMY1 CN and SAA in the context of physiological implications could be misleading because several environmental factors contribute to variability in SAA, including the time of day [19,20]. Prior studies have also reported higher salivary amylase concentration or SAA in people with T2D than in healthy controls, and it has been proposed that these measurements could be used as diagnostic biomarkers for T2D [21]. As a second goal in this study, we investigated the association between AMY1 CN and SAA in healthy individuals compared to those with self-reported T2D or prediabetes and explored the diurnal variation of SAA. Our findings implicate SAA in glucose homeostasis and provide practical insights for future research and health applications involving this gene.
Materials and methods
Participant recruitment and sample collection
The human subjects research from which the samples were derived was approved by the Cornell University Institutional Review Board (Approval Number: 1902008575 [starch study] and 1808008178 [T2D microbiome study]). The study population included two cohorts: (1) healthy individuals 18 years or older in the starch study and (2) individuals with a self-reported diagnosis of prediabetes or type 2 diabetes (T2D) 45 years or older in the T2D microbiome study. The demographics of the two cohorts are shown in S1 Table. We recruited each cohort from the local community in and around Ithaca, New York, USA. The recruitment periods for the starch study were from August 1, 2019, to December 31, 2019, and from September 2, 2020, to October 2, 2020. The recruitment period for the T2D microbiome study was from October 13, 2019, to March 14, 2020. All participants provided written informed consent before enrollment for both cohorts. To confirm the diagnosis of prediabetes or T2D for each participant in the T2D cohort, we requested a diabetes medication prescription or a doctor’s note. In addition, the participants in the T2D cohort underwent an oral mixed-meal tolerance test to assess glucose dysregulation (S1 Fig).
Approximately 5 ml of saliva was collected from the 196 participants for AMY1 CN determination. In a subset of 94 participants, up to four saliva samples were collected on different days to measure SAA. Before saliva collection, participants were asked to refrain from brushing their teeth for a minimum of six hours and to not consume any food or beverages, including water, for a minimum of 30 minutes. Saliva samples were collected using a non-invasive, self-collection method. Participants were instructed to allow saliva to passively accumulate in their mouth and then expel it into a sterile 50 ml conical tube. These samples were placed on ice immediately after collection, aliquoted within three hours, and stored at −80 °C until further analysis.
DNA extraction
Genomic DNA was extracted from saliva samples using the QIAamp DNA Blood Kit (Qiagen, cat # 51161), the QIAamp DNA Blood Mini Kit (Qiagen, cat # 51104), and the QIAamp DNA Investigator Kit (Qiagen, cat # 56504) following manufacturer protocols.
AMY1 copy number determination by quantitative PCR
We performed quantitative PCR (qPCR) using the following primers to amplify the AMY1 genes, AMY1A, AMY1B, and AMY1C: AMY1-forward: 5’-TGAGAACATTAGGCCACAGCA-3’, AMY1-reverse: 5’-TGGAAATCATCTCAATGACCTCT-3 [22]. We used EIF2B2 as a reference gene (CN = 2) and the following primers to amplify the reference gene: EIF2B2-forward: 5’-GCTCAAAGTGCTTGAGGACC-3’; EIF2B2-reverse: 5’-CAAAGCCAAACCCAGACAAT-3’. For each gene of interest, the qPCR reaction consisted of 1 μl genomic DNA (5 ng/μl), 0.5 μl of each 10 μM primer (forward and reverse), and 3 μl of PCR grade H2O, and 5 μl iTaq™ Universal SYBR® Green Supermix (Catalog # BioRad 1725122), for a total volume of 10 μl per reaction. The qPCR conditions were: initial denaturation at 95 °C for 5 minutes and 40 cycles of 95 °C for 10 seconds and 60 °C for 30 seconds on a Roche LightCycler 480 Real-Time PCR Instrument. A standard curve was made using genomic DNA NA12286 (Coriell Institute; AMY1 CN = 2). All reactions, including standards and blanks, were performed in quadruplicate. These genomic DNAs from the Coriell institute were used as positive controls on all qPCR plates to assess variability between plates: NA18972, NA12873, NA10472, NA12890, NA10852, NA12043, NA11992, NA12414, NA12340, NA06994, NA12342, NA12286, NA18522, and NA19138.
AMY1 copy number determination by droplet digital PCR
Before copy number determination by droplet digital PCR (ddPCR), we digested the genomic DNA with the HaeIII restriction enzyme (New England Biolabs) to separate the tandemly repeated copies of the AMY1 gene. This mitigates the underestimation of copy numbers in high copy number samples. Each 15 ul restriction digestion reaction consisted of 12 ul genomic DNA at 15–20 ng/ul concentration, 1.5 ul HaeIII enzyme at 2 U/ul, and 1.5 ul of 10x NEB buffer. DNA was digested for 60 minutes at 37 °C, followed by heat inactivation of the enzyme for 20 minutes at 60 °C.
We determined AMY1 CN compared to a reference gene AP3B1 using ddPCR on the Bio-Rad QX100 AutoDG Droplet Digital PCR System. The PCR amplification reaction included 10 ul of 2x ddPCR Supermix for Probes No dUTP (Bio-Rad), 2 ul of restriction enzyme digested genomic DNA at a concentration of 12–16 ng/ul, 0.5 ul of AMY1 TaqMan assay (Hs07226361_cn, ThermoFisher Scientific), and 0.5 ul of reference gene AP3B1 assay (dHsaCP1000001, Bio-Rad) for a total reaction volume of 20 ul. The reaction was emulsified with Droplet Generator Oil (Bio-Rad) using the QX100 Droplet Generator (Bio-Rad) per the manufacturer’s instructions. AMY1 CN was determined using QuantaSoft software version 1.7.4.0917 (Bio-Rad laboratories).
Salivary amylase activity assay
We used the Salimetrics Salivary Alpha-Amylase Enzymatic Kit (Salimetrics, cat 1–1902) to measure SAA. We performed the assay in triplicate for each saliva sample. We followed the manufacturer protocol except for using 300 ul amylase substrate per reaction instead of 320 ul. The calculations were adjusted according to the change in total assay volume. The triplicate measurements for each participant were averaged to obtain the mean SAA to be used in the analyses.
Statistical analysis
For all participants, we determined the AMY1 CN with qPCR and ddPCR in two separate experiments. For statistical analyses comparing qPCR and ddPCR, we calculated the mean CN value for each method by averaging the results of the two separate experiments. For all the analyses examining the association of AMY1 CN with other parameters, we used the mean CN value obtained from the two ddPCR experiments for each individual.
All statistical analyses were conducted using Rstudio (4.2.2), and a p-value of ≤0.05 was considered statistically significant. We fit linear mixed models to examine how SAA is influenced by AMY1 CN, the time of saliva sample collection, and T2D status. The response variable, SAA, was log-transformed to meet the assumptions of linear models, and the normality of the model residuals was assessed. We included subject ID as a random effect because multiple samples were collected from each participant at different times. The linear mixed models were conducted with the lme4 package in R [23]. Figures depicting the data points used in the linear models also show the fitted lines of the model-predicted values. We also used linear regression to examine the relationship between AMY1 CN values determined using qPCR and ddPCR and used intra-class correlation (ICC) to check the consistency between qPCR and ddPCR values. S2–S7 Tables show the R output for each of the models.
Results
qPCR and ddPCR estimates for AMY1 CN
The median AMY1 CN obtained with qPCR was 7.5 with a range of 2–19 copies, and the median AMY1 CN obtained with ddPCR was 7 with a range of 2–20 copies. We compared the AMY1 CN estimates of the reference Coriell DNAs obtained using qPCR and ddPCR in our lab to the values previously reported in the literature (Table 1). We calculated the total AMY1 CN obtained using QuicK-mer by adding all copies of the salivary amylase genes (AMY1A, AMY1B, AMY1C) in S1 Table of Shen et al. [25]. For each Coriell DNA, the standard deviation of all the AMY1 CN estimates was between 0.20–2.54 with a median standard deviation of 0.73 when we exclude NA10472 and NA10852. The greatest standard deviation in CN estimates was for NA18972, which has the highest AMY1 CN.
We used linear regression to examine the association between the CNs determined by qPCR and ddPCR for all 210 samples, which included 196 study participants and the Coriell DNAs. qPCR-estimated AMY1 CNs were used to predict ddPCR-estimated CNs (S2 Table). qPCR measurements were predictive of the ddPCR measurements (β = 0.917, p < 0.001). However, the qPCR-estimated CN values were more dispersed (Fig 1), suggesting that for a given AMY1 CN, qPCR produced more variable estimates than ddPCR. The variability in qPCR estimates for given ddPCR estimates of AMY1 CN may indicate lower precision in qPCR, even though the overall variance between the two methods was not statistically different, as shown by the F-test (F = 1.07, df = 209, p = 0.63). To further assess agreement between the two techniques, we calculated the intraclass correlation coefficient (ICC; S3 Table). The ICC coefficient was 0.95 (95% CI: 0.93–0.96), indicating a high level of consistency between qPCR and ddPCR in estimating AMY1 CN.
The values on the x-axis and y-axis are the mean CN of two replicates of qPCR and ddPCR, respectively. For each qPCR replicate, the reactions were run in quadruplicate. The plot shows individual CN estimates with a linear regression line.
Saliva sample collection time
A linear mixed model was used to examine the diurnal pattern of SAA, with log-transformed SAA as the response variable and time of saliva collection as the predictor (S4 Table). Participant was included as a random effect to account for interindividual variability. The results revealed diurnal variation in SAA—SAA increased as the day progressed (n = 263 samples; β = 0.09; p < 0.001). A previous study reported a sharp increase in SAA from morning to noon followed by a slower rate of change throughout the afternoon [26]. To evaluate whether this pattern occurred in our dataset, we ran the model separately for morning (t ≤ 12:00) and afternoon (t > 12:00) samples, to compare the rate of increase in SAA between the two periods. We observed that SAA increased significantly with time in the morning (n = 111 samples; β = 0.18; p = 0.003), and after noon (n = 152 samples; β = 0.085; p = 0.05). For every one hour increase in the time of collection, the expected value of SAA increased by 19% within the samples collected at or before noon. The percent increase was 9% in the samples collected after noon.
To determine whether the difference in the rate of increase in SAA between morning and afternoon was significant, we used a linear mixed model to predict the log of the SAA by the time of saliva collection during the day, the time of day (categorical as morning/afternoon), and their interaction. We also included a random effect of participant to account for repeat measurements from participants (S5 Table). There was no significant difference in the rate of increase in SAA with time between morning and afternoon (β = 0.107; p = 0.1) (Fig 2).
The data points in green represent samples collected at or before noon (n = 111). The data points in purple represent the samples collected after noon (n = 152). The solid lines represent predicted values from the linear mixed model fitted for samples collected at or before noon and after noon. The rate of change of SAA does not significantly differ between morning and afternoon (p = 0.1).
AMY1 CN and SAA
We used a linear mixed effects model with log-transformed SAA as the response variable, AMY1 CN obtained by ddPCR as a fixed effect, and participant as a random effect to investigate the association between AMY1 CN and SAA (S6 Table). We found a significant positive association between AMY1 CN and SAA (n = 300 samples; β = 0.15; p < 0.001). We assessed model fit using marginal R² and conditional R² in the rsq package [27] in R. AMY1 CN alone explained 18% of the variance in SAA. There was a high conditional R2 (75%) indicating that AMY1 CN and interindividual differences, due to potential biological or other factors not included in the model, explain most of the variance in SAA.
T2D/prediabetes status modifies the relationship between AMY1 CN and SAA
To evaluate whether the association between AMY1 CN and SAA varied in individuals with T2D or prediabetes compared to healthy controls, we fit a linear mixed effects model with the log of SAA as a response variable, and with AMY1 CN, health status (T2D/prediabetes or healthy control), age, sex, and time of saliva collection as predictors (S7 Table). An interaction term between AMY1 CN and T2D/prediabetes status was included. We excluded individuals with missing data for any variables in the model. The interaction between AMY1 CN and T2D/prediabetes status was significant (n = 285 samples; β = 0.22; p = 0.006). For each additional copy of AMY1, SAA increased by approximately 14% in the control group and 43% in the T2D/prediabetes group. Taken together, with increasing AMY1 CN, SAA was higher for each additional copy of AMY1 CN in individuals with T2D or prediabetes compared to controls (Fig 3). These findings suggest that T2D/prediabetes status modifies the relationship between AMY1 CN and SAA.
The data points are colored according to the participants’ T2D/prediabetes status with up to four samples collected per individual. For T2D/prediabetes, blue data points n = 27; for healthy controls, yellow data points n = 258. The lines in this figure represent predicted values from the linear mixed model. For the interaction between AMY1 CN and T2D/prediabetes status, p = 0.006.
Discussion
The key finding in our study is that T2D/prediabetes status modifies the association between AMY1 CN and SAA. In addition, we found evidence to support diurnal variations in SAA. We observed a positive correlation between SAA and the time of day at which saliva samples were collected. We also showed that qPCR values can be comparable to ddPCR values when measuring AMY1 CN.
Whereas AMY1 CN is a fixed measure, fluctuations in SAA could provide insight about the day-to-day impact of salivary amylase on blood glucose homeostasis. We observed that compared to healthy controls, people with prediabetes or T2D had a greater increase in SAA with an increase in AMY1 CN. At least four other studies found that T2D patients had elevated SAA compared to controls [28–31]. Our analysis, confirming higher SAA in T2D/prediabetes patients, included samples collected from T2D/prediabetes patients and controls in the morning and after noon. Our statistical model included time of collection as a covariate. Since participants with T2D or prediabetes have aberrant glucose homeostasis, these findings suggest that higher SAA is associated with glucose dysregulation. If we assume that participants with T2D/prediabetes had lower SAA before developing the glucose dysregulation that resulted in their diagnosis, an increase in SAA from one’s previous mean activity may be indicative that disease onset is pending. Therefore, increased SAA could either be contributing to the development of T2D/prediabetes or counteracting glucose dysregulation.
Most studies assessing the long-term effects of salivary amylase dosage and glucose metabolism are reported in terms of AMY1 CN, as opposed to SAA. Two studies have reported that a higher AMY1 CN is associated with an increased postprandial glucose response to starches, with both studies also reporting a positive correlation between AMY1 CN and SAA [7,32]. This result suggests that a higher CN, with presumably higher SAA, would more efficiently liberate sugars from starchy foods to raise blood glucose. When attempting to minimize steep increases in postprandial glucose response to dietary intake for the prevention or management of T2D, a high AMY1 CN could be a challenge for people with a high starch intake. However, one study reported an inverse association between AMY1 CN and fasting blood glucose in people with a high starch intake [32]. This suggests there is some protective mechanism stabilizing blood glucose long term in people with high AMY1 CN. Two studies reported that high SAA is associated with low glucose response to starch in a small number of subjects [33,34]. Mandel and Breslin posited that this is mediated by preabsorptive insulin release, which was observed in high SAA participants [34]. This would support the premise of a protective mechanism with increased SAA being a compensatory effect blunting blood glucose response to starch ingestion. Congruent with the premise that high AMY1 CN has a protective effect, there are several reports that high AMY1 CN is associated with decreased insulin resistance and T2D risk [8,12]. Additionally, another large study reported an inverse association between starch intake and T2D in individuals with AMY1 CN ≥ 10 [5].
One limitation is that our study was cross-sectional. Since we only measured SAA in each individual during their current disease status—when they were either with or without prediabetes or T2D—we do not know whether the T2D/prediabetes participants had higher SAA throughout their adult lives, or their SAA increased before or after disease onset. A potential limitation of our findings involving SAA and T2D status is the relative sample sizes of the groups—there were 16 individuals with T2D or prediabetes and 74 individuals without T2D or prediabetes included in the analysis. Another potential limitation is the difference in the age range of the healthy cohort (18–59 years) and the T2D/ prediabetes cohort (47–87 years). However, when we performed a linear regression analysis to investigate the relationship between age and SAA, we found no significant association between the two variables. Furthermore, SAA does not appear to vary as adults age, although the studies addressing this question were cross-sectional [35,36]. A study that would best address the question of whether SAA could be a biomarker for the onset of T2D would be prospective in several populations of different ethnic groups and with multiple risk factors for T2D. This should include AMY1 CN determination and measurements of SAA and glucose response to starch at multiple time points.
We also demonstrated that qPCR can be a reliable alternative for copy number assessment with comparable accuracy to ddPCR, which has been proposed as the model method for quantifying a target gene copy number [37]. This finding is not consistent with another study that reported discordance between qPCR and ddPCR values for AMY1 CN in saliva samples [38]. Advantages of ddPCR include sensitivity and no requirement for a standard curve. Conversely, qPCR is less expensive and faster than ddPCR and, thus, more adaptable for high throughput protocols with automation. Another consideration, when using ddPCR for quantifying a target gene CN, is the method poses an inadvertent risk of underestimating CNs when multiple copies of the target gene are linked in tandem [39]. As previous groups have done, we addressed this by initially digesting the genomic DNA using a restriction enzyme. This additional step in the protocol augments the processing time and expense. Thus, qPCR may be preferred over ddPCR for AMY1 CN estimation if there are a large number of samples being genotyped or there is restricted access to resources.
We pose a question regarding the interchangeability of AMY1 CN and SAA when studying metabolic health. AMY1 CN remains stable within an individual through their lifetime. This genetic stability could make it a reliable metric to assess an individual’s metabolic predisposition. Unlike AMY1 CN, SAA fluctuates and can be affected by several factors, e.g., time of sample collection, caffeine intake, smoking, and stress, in addition to AMY1 CN [35]. We observed that, although SAA positively correlates with AMY1 CN, there is diurnal variation. The pattern in SAA is a rapid increase in the morning followed by a slower increase in the afternoon, consistent with previous findings [26]. Further investigation is required to determine the impact of these differences in diurnal patterns on metabolic health. A key implication is that consistent timing of saliva collection across subjects would be ideal when investigating the role of SAA in health and disease.
Supporting information
S1 Fig. Mixed meal tolerance test in participants with type 2 diabetes or prediabetes.
Plot illustrating the plasma glucose and insulin responses during a mixed meal tolerance test at 0, 15, and 30 minutes. The figure displays three diabetes status groups: orange–individual with no T2D/prediabetes (n = 1); green–individuals with prediabetes (n = 4); blue–individuals with T2D (n = 13). Mixed meal tolerance test data were not available for one participant with T2D.
https://doi.org/10.1371/journal.pone.0324660.s001
(DOCX)
S1 Table. Table comparing individuals with T2D/prediabetes to those without T2D/prediabetes (controls).
This table shows the mean and standard deviation of the participant demographics. The healthy controls (n = 76) included individuals who submitted up to four saliva samples at different times of the day. The T2D/prediabetes group (n = 18) includes individuals who submitted up to two saliva samples.
https://doi.org/10.1371/journal.pone.0324660.s002
(DOCX)
S2 Table. R output for linear regression to test the association between AMY1 copy number values determined by qPCR and ddPCR.
https://doi.org/10.1371/journal.pone.0324660.s003
(DOCX)
S3 Table. R output for the intraclass correlation coefficient (ICC) to assess consistency between qPCR and ddPCR.
https://doi.org/10.1371/journal.pone.0324660.s004
(DOCX)
S4 Table. A) R output for linear mixed regression testing association between SAA and the time of saliva collection during the day using samples collected in the morning and in the afternoon.
B, C) R output for linear mixed models testing association between SAA and time of saliva sample collection in B) morning samples only and C) afternoon samples only.
https://doi.org/10.1371/journal.pone.0324660.s005
(DOCX)
S5 Table. R output for linear mixed regression testing for the difference between rates of SAA change in the morning and the afternoon.
https://doi.org/10.1371/journal.pone.0324660.s006
(DOCX)
S6 Table. R output for linear mixed regression model testing for association between SAA and AMY1 CN.
https://doi.org/10.1371/journal.pone.0324660.s007
(DOCX)
S7 Table. R output for linear mixed regression model assessing the effect of type 2 diabetes status on the association between AMY1 copy number and salivary amylase activity.
https://doi.org/10.1371/journal.pone.0324660.s008
(DOCX)
Acknowledgments
We thank Erika Mudrak from the Cornell Statistical Consulting Unit for assistance with the analyses and Dorothy Kim Superdock for her help with sample collection and processing. We used DNA from human lymphoblastoid cell lines from the NHGRI Sample Repository for Human Genetic Research and the NIGMS Human Genetic Cell Repository of the Coriell Institute for Medical Research. We thank the Biotechnology Resource Center Genomics Facility at the Cornell Institute of Biotechnology for performing ddPCR.
References
- 1. Zarrei M, MacDonald JR, Merico D, Scherer SW. A copy number variation map of the human genome. Nat Rev Genet. 2015;16(3):172–83. pmid:25645873
- 2. Chaignat E, Yahya-Graison EA, Henrichsen CN, Chrast J, Schütz F, Pradervand S, et al. Copy number variation modifies expression time courses. Genome Res. 2011;21(1):106–13. pmid:21084671
- 3. Rice AM, McLysaght A. Dosage-sensitive genes in evolution and disease. BMC Biol. 2017;15(1):78. pmid:28863777
- 4. Bolognini D, Halgren A, Lou RN, Raveane A, Rocha JL, Guarracino A., et al. Recurrent evolution and selection shape structural diversity at the amylase locus. Nature. 2024. Oct;634(8034):617-625. pmid:39232174
- 5. Hamid AK, Andersson-Assarsson J, Ericson U, Sonestedt E. Interaction effect between copy number variation in Salivary Amylase Locus (AMY1) and starch intake on glucose homeostasis in the Malmö Diet and Cancer cohort. Front Nutr. 2021;7:598850. pmid:33490099
- 6. Rukh G, Ericson U, Andersson-Assarsson J, Orho-Melander M, Sonestedt E. Dietary starch intake modifies the relation between copy number variation in the salivary amylase gene and BMI. Am J Clin Nutr. 2017;106(1):256–62. pmid:28539377
- 7. Atkinson FS, Hancock D, Petocz P, Brand-Miller JC. The physiologic and phenotypic significance of variation in human amylase gene copy number. Am J Clin Nutr. 2018;108(4):737–48. pmid:30239565
- 8. Liu Y, Smith CE, Parnell LD, Lee YC, An P, Straka RJ., et al. Salivary AMY1 copy number variation modifies age-related type 2 diabetes risk. Clin Chem. 2020;66(5):718–26. pmid:32337541
- 9. Carpenter D, Dhar S, Mitchell LM, Fu B, Tyson J, Shwan NAA., et al. Obesity, starch digestion and amylase: association between copy number variants at human salivary (AMY1) and pancreatic (AMY2) amylase genes. Hum Mol Genet. 2015;24(12):3472–80. pmid:25788522
- 10. Usher CL, Handsaker RE, Esko T, Tuke MA, Weedon MN, Hastie AR, et al. Structural forms of the human amylase locus and their relationships to SNPs, haplotypes and obesity. Nat Genet. 2015;47(8):921–5. pmid:26098870
- 11. Falchi M, El-Sayed Moustafa JS, Takousis P, Pesce F, Bonnefond A, Andersson-Assarsson JC, et al. Low copy number of the salivary amylase gene predisposes to obesity. Nat Genet. 2014;46(5):492–7. pmid:24686848
- 12. Choi Y-J, Nam Y-S, Yun JM, Park JH, Cho BL, Son H-Y, et al. Association between salivary amylase (AMY1) gene copy numbers and insulin resistance in asymptomatic Korean men. Diabet Med. 2015;32(12):1588–95. pmid:25996848
- 13. Hasegawa T, Kakuta M, Yamaguchi R, Sato N, Mikami T, Murashita K., et al. Impact of salivary and pancreatic amylase gene copy numbers on diabetes, obesity, and functional profiles of microbiome in Northern Japanese population. Sci Rep. 2022;12:7628. pmid:35538098
- 14. Perry GH, Dominy NJ, Claw KG, Lee AS, Fiegler H, Redon R, et al. Diet and the evolution of human amylase gene copy number variation. Nat Genet. 2007;39(10):1256–60. pmid:17828263
- 15. Marquina C, Mousa A, Belski R, Banaharis H, Naderpoor N, De Courten B. Increased inflammation and cardiometabolic risk in individuals with low AMY1 copy numbers. J Clin Med. 2019;8(3):382. pmid:30893908
- 16. Shwan NAA, Armour JAL. No evidence for association of BMI with salivary amylase gene copy number in the UK 1958 birth cohort. Obesity (Silver Spring). 2019;27(9):1533–8. pmid:31441234
- 17. Valsesia A, Kulkarni SS, Marquis J, Leone P, Mironova P, Walter O, et al. Salivary α-amylase copy number is not associated with weight trajectories and glycemic improvements following clinical weight loss: results from a 2-phase dietary intervention study. Am J Clin Nutr. 2019;109(4):1029–37. pmid:30982860
- 18. White S. Counting copy number and calories. Nat Genet. 2015;47(8):852–3. pmid:26220133
- 19. Gachons CP, Breslin PAS. Salivary amylase: digestion and metabolic syndrome. Curr Diab Rep. 2016;16(10):102. pmid:27640169
- 20. Nater UM, Rohleder N, Schlotz W, Ehlert U, Kirschbaum C. Determinants of the diurnal course of salivary alpha-amylase. Psychoneuroendocrinology. 2007;32(4):392–401. pmid:17418498
- 21. Pérez-Ros P, Navarro-Flores E, Julián-Rochina I, Martínez-Arnau FM, Cauli O. Changes in salivary amylase and glucose in diabetes: a scoping review. Diagn. (Basel) 2021;11(3):453. pmid:33800850
- 22. Poole AC, Goodrich JK, Youngblut ND, Luque GG, Ruaud A, Sutter JL. Human salivary amylase gene copy number impacts oral and gut microbiomes. Cell Host Microbe. 2019;25(4):553-564.e7. pmid:30974084 https://doi.org/10.1016/j.chom.2019.01.001
- 23. Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Softw. 2015;67:1–48.
- 24. Carpenter D, Mitchell LM, Armour JAL. Copy number variation of human AMY1 is a minor contributor to variation in salivary amylase expression and activity. Hum Genomics. 2017;11(1):2. pmid:28219410
- 25. Shen F, Kidd JM. Rapid, paralog-sensitive CNV analysis of 2457 human genomes using QuicK-mer2. Genes (Basel). 2020;11(2):141. pmid:32013076
- 26. Harmon AG, Towe-Goodman NR, Fortunato CK, Granger DA. Differences in saliva collection location and disparities in baseline and diurnal rhythms of alpha-amylase: a preliminary note of caution. Horm Behav. 2008;54(5):592–6. pmid:18640119
- 27. Dabao Z. rsq: R-squared and related measures. Accessed 2023 October 1. https://CRAN.R-project.org/package=rsq
- 28. Artino M, Dragomir M, Ionescu S, Bădiţa D, Niţă V, Chiţoi E. Diurnal behaviour of some salivary parameters in patients with diabetes mellitus (protein concentration, amylase activity, density)--note I. Rom J Physiol. 1998;35(1–2):79–84. pmid:11000868
- 29. Shah VS, Pareikh D, Manjunatha BS. Salivary alpha-amylase-biomarker for monitoring type II diabetes. J Oral Maxillofac Pathol. 2021;25(3):441–5. pmid:35281155
- 30. Abd-Elraheem SE, El Saeed AM, Mansour HH. Salivary changes in type 2 diabetic patients. Diabetes Metab Syndr. 2017;11(Suppl 2):S637-41. pmid:28511885
- 31. Dodds MW, Dodds AP. Effects of glycemic control on saliva flow rates and protein composition in non-insulin-dependent diabetes mellitus. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 1997;83(4):465–70. pmid:9127379
- 32. Farrell M, Ramne S, Gouinguenet P, Brunkwall L, Ericson U, Raben A, et al. Effect of AMY1 copy number variation and various doses of starch intake on glucose homeostasis: data from a cross-sectional observational study and a crossover meal study. Genes Nutr. 2021;16(1):21. pmid:34789141
- 33. Barling PM, Shyam S, Selvathevan MD, Misra S. Anomalous association of salivary amylase secretion with the postprandial glycaemic response to starch. BMC Nutrition. 2016;2:50. https://doi.org/10.1186/s40795-016-0088-0
- 34. Mandel AL, Breslin PAS. High endogenous salivary amylase activity is associated with improved glycemic homeostasis following starch ingestion in adults. J Nutr. 2012;142(5):853–8. pmid:22492122
- 35. Rohleder N, Nater UM. Determinants of salivary α-amylase in humans and methodological considerations. Psychoneuroendocrinology. 2009;34(4):469–85. pmid:19155141
- 36. Nassar M, Hiraishi N, Islam MS, Otsuki M, Tagami J. Age-related changes in salivary biomarkers. J Dent Sci. 2014;9(1):85–90. https://doi.org/10.1016/j.jds.2013.11.002
- 37. Pinheiro LB, Coleman VA, Hindson CM, Herrmann J, Hindson BJ, Bhat S., et al. Evaluation of a droplet digital polymerase chain reaction format for DNA copy number quantification. Anal Chem. 2012;84(2):1003–11. pmid:22122760
- 38. Ooi DSQ, Tan VMH, Ong SG, Chan YH, Heng CK, Lee YS. Differences in AMY1 gene copy numbers derived from blood, buccal cells and saliva using quantitative and droplet digital PCR methods: flagging the pitfall. PLoS One. 2017;12(1):e0170767. pmid:28125683
- 39. Bell AD, Usher CL, McCarroll SA. Analyzing copy number variation with droplet digital PCR. Methods Mol Biol. 2018;1768:143–60. pmid:29717442