Search in the Abstract Database

Abstracts Search 2019

P822 Genome-wide association study (GWAS) of a Maltese inflammatory bowel disease cohort

J. Schembri*1, N. Pace2, S. Vella1, N. Piscopo1, F. Degenhardt3, A. Franke3, P. Ellul1

1Mater Dei Hospital, Gastroenterology, Msida, Malta, 2University of Malta, Department of Biochemistry, Msida, Malta, 3Christian-Albrechts-University of Kiel, Institute of Clinical Molecular Biology, Kiel, Germany


Whilst most of the early inflammatory bowel disease (IBD) genetic studies were performed on Caucasian subjects, it is now increasingly recognised that different genes might be involved in different populations, especially from different ethnic backgrounds. Prior research in Maltese IBD patients had in fact determined that prevalence of NOD2 polymorphisms was very low, in stark contrast to IBD patients from mainland Europe. Hence, the aim of this study was to genotypically characterise a discovery cohort consisting of Maltese IBD patients.


We conducted a case–control genetic association study using a hypothesis-free approach. Genotyping was carried out on the Illumina Immunochip platform.


After strict quality control 93 ulcerative colitis (UC), 160 Crohn's (CD) patients and 188 healthy controls remained. Figure 1 demonstrates no significant population stratification with evidence of several disease associated loci.

Figure 1. QQ-plot of the entire QCed IBD study population. The graph is linear, in keeping with no significant population stratification. Compelling evidence for an excess of disease associations is represented by the tail end.

Separate analyses were carried out for CD and UC and these are reported in Tables 1 and 2, respectively. Disease association with CD were larger in number and strength, compared with UC.

Table 1. Genome-wide association results from a cohort of Maltese CD patients in descending order of significance. Two SNPs on chromosome 16 (CDH1 gene) demonstrate strong linkage disequilibrium.

Table 2. Genome-wide association results from a cohort of Maltese UC patients in descending order of significance. Weaker associations have emerged from this analysis possibly related to the smaller sample size.

Furthermore, whilst some loci contributed to both types of IBD, most demonstrated preferential association with either CD or UC (Figure 2).

Figure 2. Odds ratio (OR) plot showing IBD-type specificity for typed SNPs. The 53 independent signals, plotted by total IBD OR and phenotype specificity (measured by the OR of CD relative to UC), have been coloured according to IBD phenotype.


This study is the first genotype-association study in the Maltese IBD population and despite the relatively small sample size and lack of a replication cohort we recruited approximately one fourth of the country's entire IBD population. RAD51B, CDH1, HLA-DPB2 and MS4A5 represented the most significant polymorphisms and whilst several loci in our cohort are amongst the 231 known IBD risk loci, several others have never been associated with IBD. The previously reported low-prevalence of NOD2 polymorphisms in Maltese patients has been confirmed in our cohort, further validating our findings and highlighting the importance of biomedical research in small populations and under-represented communities.