Inference of causal relationships between sleep-related traits and 1,527 phenotypes using genetic data

Our plain-English summary of Inference of causal relationships between sleep-related traits and 1,527 phenotypes using genetic data, by L.M. García-Marín, A.I. Campos, N.G. Martin, et al (Sleep, 2021).

The study

• Authors: L.M. García-Marín, A.I. Campos, N.G. Martin, G. Cuéllar-Partida, M.E. Rentería
• Journal: Sleep
• Published: 2021;44(1)
• Cited by: 15 articles, as counted by NIH iCite
• Read it: PubMed · Publisher

Abstract

Study Objective: Sleep is essential for both physical and mental health, and there is a growing interest in understanding how different factors shape individual variation in sleep duration, quality and patterns, or confer risk for sleep disorders. The present study aimed to identify novel inferred causal relationships between sleep-related traits and other phenotypes, using a genetics-driven hypothesis-free approach not requiring longitudinal data.

Methods: We used summary-level statistics from genome-wide association studies and the latent causal variable (LCV) method to screen the phenome and infer causal relationships between seven sleep-related traits (insomnia, daytime dozing, easiness of getting up in the morning, snoring, sleep duration, napping, and morningness) and 1,527 other phenotypes.

Results: We identify 84 inferred causal relationships. Among other findings, connective tissue disorders increase insomnia risk and reduce sleep duration; depression-related traits increase insomnia and daytime dozing; insomnia, napping, and snoring are affected by obesity and cardiometabolic traits and diseases; and working with asbestos, thinner, or glues may increase insomnia risk, possibly through an increased risk of respiratory disease or socio-economic related factors.

Conclusion: Overall, our results indicate that changes in sleep variables are predominantly the consequence, rather than the cause, of other underlying phenotypes and diseases. These insights could inform the design of future epidemiological and interventional studies in sleep medicine and research.

© Sleep Research Society 2020. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail [email protected].

In plain English

This study used published genetic data rather than examining people directly. The authors took genome-wide association results for seven self-reported sleep traits in people of European ancestry. Using a statistical method called latent causal variable analysis, they compared them with 1,527 other traits, mostly from the UK Biobank. The method estimates whether a shared genetic link between two traits suggests one influences the other. Across the sleep traits they identified 84 inferred causal relationships. Insomnia was genetically correlated with 608 traits, and 48 showed an inferred causal effect on insomnia risk. Among the strongest was the questionnaire item "often worked with materials containing asbestos" (genetic correlation 0.26, genetic causality proportion -0.97). "Rarely/never worked" with such materials was associated with reduced insomnia risk. Interstitial lung disease and COPD also showed inferred effects on insomnia. The authors suggested that exposure to asbestos and solvents could lead to insomnia, perhaps through severe respiratory disease. They recommended treating the method as a hypothesis generator, and noted that the data came only from European-ancestry UK Biobank participants.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Genetic study using published genome-wide association study (GWAS) summary statistics and latent causal variable (LCV) analysis to screen for inferred causal relationships, with Mendelian randomisation as a sensitivity analysis
• Who was studied: GWAS summary statistics from samples of European ancestry for seven self-reported sleep traits (average sample sizes from 344,084 for morningness to 408,317 for snoring; 384,891 for insomnia), compared with 1,527 other GWAS, most from the UK Biobank (Neale Lab)
• Exposure: Among the 1,527 traits screened were UK Biobank work-environment items, including "often worked with materials containing asbestos" and "rarely/never worked with materials containing asbestos". The article does not name fiber types, products or doses.
• Main results:
  • Insomnia was significantly genetically correlated with 608 traits. 48 of these showed an inferred causal effect on insomnia risk, and no causal effect of insomnia on another trait was found.
  • Often worked with materials containing asbestos was among the traits with the strongest evidence of a causal effect on insomnia: rG = 0.26, GCP = -0.97, p-value for GCP = 7.23 x 10^-202. Table 2 prints rG as 0.266.
  • Rarely/never worked with materials containing asbestos was associated with reduced insomnia risk: rG = -0.44, GCP = -0.92, p-value for GCP = 2.30 x 10^-33.
  • Interstitial lung disease and COPD each showed evidence of increasing insomnia risk (text: rG = 0.38, GCP = -0.65, p-value for GCP = 0.003).
  • Other workplace items, such as a workplace often full of chemicals or other fumes and often having a lot of diesel exhaust, were genetically correlated with insomnia but did not show an inferred causal association. No inferred causal association was seen for the Townsend deprivation index.
  • Across all seven sleep traits, 84 inferred causal relationships were identified. Only 11 could be tested with traditional Mendelian randomisation methods. The text does not say whether the asbestos relationship was among those 11.
• The authors' conclusion: The authors concluded that changes in sleep variables were mainly the consequence rather than the cause of other conditions. They suggested that exposure to asbestos and solvents could lead to insomnia, perhaps as a consequence of severe respiratory diseases such as COPD, ILD and asthma.
• Limitations the authors state:
  • Only data from individuals of European ancestry from the UK Biobank were used, which may limit generalizability.
  • The LCV method assumes no bidirectional causality and no confounding by environmental correlates of genotypes.
  • Estimates depend on the statistical power of the underlying GWAS, which limits the ability to detect effects for some traits.
  • The analysis is bivariate, so a third trait sharing genetic components could explain an association between two traits.
  • Most LCV findings could not be checked with Mendelian randomisation, partly because the requirement for independent samples could not be met.
  • The authors recommend interpreting LCV results as hypotheses to be followed up with other methods.

Related studies

• Toxicology and epidemiology: improving the science with a framework for combining toxicological and epidemiological evidence to establish causal inference — H.O. Adami, S.C. Berry, et al. (2011)
• A quantitative weight of evidence assessment of Hill's guidelines for causal inference for cosmetic talc as a cause of mesothelioma — A.M. Ierardi, A. Urban, et al. (2021)
• Year in review 2011: acute lung injury, interstitial lung diseases, physiology, sleep and lung cancer — K. Takahashi, N.D. Eves, et al. (2012)
• Bidirectional Mendelian randomization and potential mechanistic insights into the causal relationship between gut microbiota and malignant mesothelioma — Y. Zhou, H. Zhu, et al. (2025)
• Tracheal carcinoma and mixed pneumoconiosis. A causal relationship? — D. Theegarten, S. Philippou, et al. (1995)
• Genome-wide profile of pleural mesothelioma versus parietal and visceral pleura: the emerging gene portrait of the mesothelioma phenotype — O.D. Røe, E. Anderssen, et al. (2009)

Browse return to medical research.