Genetic causes and phenotypic consequences of breastfeeding

PhD project (3/4 yr research project leading to independent research at the doctorate level)

Dr Luisa Zuccolo, Prof Richard Martin, Prof George Davey Smith

Return to list


Breastfeeding is recommended as the best source of nutrition in early life by the WHO, with claims of numerous benefits for both mother and child both in the short and long term. However, questions remain about whether these are real (causal) benefits, and about their importance(1). One way to provide more robust evidence about the effects of breastfeeding could be to use an alternative analytical approach known as Mendelian randomization(2). This relies on testing the association between the genetic predisposition to breastfeed (theoretically independent from any other factor), and the outcome of interest, so that the resulting estimate is not affected by confounding in the way that the effect estimated through conventional multivariable regression would be.
In order to conduct MR for breastfeeding, genome-wide association studies are first needed to uncover common genetic variation underpinning breastfeeding duration, which seems highly heritable(3).

Aims & objectives

1. To identify common genetic variants predictive of breastfeeding initiation and duration through GWAS
2. To derive polygenic risk scores (PGRS) for both traits, and predict the proportion of the variance explained in the traits
3. To use the main GWAS hits and the PGRSs as instrumental variables in Mendelian randomization analyses looking at the causal effect of breastfeeding on offspring and maternal outcomes


The studentship will involve a number of different methodologies and the use of data of various nature already collected as part of large population-based prospective studies such as the Norwegian HUNT and MoBa study and the Bristol-based Avon Longitudinal Study of Parents and Children. The student will have the opportunity to familiarise with different research methodologies. These will include analysis of genomic data in relation to breastfeeding phenotypes (GWAS, GCTA, the creation of PGRS…), and causal analysis methods (Mendelian randomization). The three independent cohorts, with participants recruited in different countries and over a period of 40+ years, will allow replication work (to rule out the effect of chance findings) and cross-cohort comparisons (to rule out the effect of biases likely to affect one study setting but not all in equal measure).


1. Victora, Cesar G et al. The Lancet , Volume 387 , Issue 10017 , 475 - 490
2. Davey Smith G and Ebrahim S, International Journal of Epidemiology 2003;32:1-22
3. Colodro-Conde L et al, Twin Res Hum Genet. 2015 Feb; 18(1): 61–72.

Created on Nov. 14, 2016, 3:47 p.m.