Multivariate and Gene-Based Association Testing of Sarcopenia: Bushehr Elderly Health Program (BEH)
Abstract
Introduction: There is a strong correlation between the skeletal muscle mass index (SMI) and handgrip strength as indicators of sarcopenias. Multivariate methods can be exploited statistical power in determining the association between these correlated heritable indicators.
Methods: We conducted a multivariate candidate-gene study based on data collected from the ongoing Bushehr Elderly Health (BEH) cohort, which evaluated the prevalence of musculoskeletal disorders in 2772 Iranians over 60 years old with 663377 single nucleotide polymorphisms (SNPs). We chose genetic variants on IL10 (chromosome 1: 206940947, 206945839), a strongly associated gene known to cause muscle diseases, as candidate regions, which included 27 independent SNPs with LD<0.4 (MAF>0.01 and p-valuehwe >0.05). MultiPhen uses a linear combination of genotypes, including SMI and handgrip, to obtain stronger statistical power. To outperform and confirm the MultiPhen results, it combined with a summary statistics level genebased association test, GATES.
Results: Among the participants, 1138 men (48%) and 1205 women (52%) aged 69.2±6.35 and 69.56±6.45, were present respectively. 27 SNPs with a maximum MAF of 0.488 and a minimum of 0.0098, p-value hwe=0.3 were selected on Interleukin 10 (IL10). In the joint model MultiPhen test, 3 intronic variants (rs11119603, rs3950619, rs57461190) were associated with IL10 with effect sizes between 0.178 and 0.883 (p-value<0.05). We used the GATES model to assess the multivariate aggregated effect of IL10 on the phenotypes. Using this method, the gene's effect was significant (0.046), showing that it is a risk gene for sarcopenia.
Conclusion: This study examined the association of handgrip, SMI, with IL10, as demonstrated in previous studies as risk factors for muscular diseases, using multivariate methods that utilized a joint model to achieve a high level of statistical power.