Genetic linkage of mild malaria to the major histocompatibility complex in Gambian children: study of affected sibling pairs

BMJ 1997; 315 doi: (Published 12 July 1997) Cite this as: BMJ 1997;315:96
  1. Annette Jepson, Wellcome Trust career development fellowa,
  2. Fatoumatta Sisay-Joof, higher scientific officerb,
  3. Winston Banya, statisticianb,
  4. Musa Hassan-King, scientistb,
  5. Angela Frodsham, research assistanta,
  6. Stephen Bennett, senior lecturerc,
  7. Adrian Hill, professor of human geneticsa,
  8. Hilton Whittle, deputy directorb
  1. a Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford OX3 7BN
  2. b Medical Research Council Laboratories, Fajara, The Gambia
  3. c Tropical Health Epidemiology Unit, London School of Hygiene and Tropical Medicine, London WC1E 7HT
  1. Correspondence to: Dr Jepson
  • Accepted 12 December 1996


Case-control studies have shown that genes for the major histocompatibility complex influence the presentation and outcome of severe Plasmodium falciparum disease (cerebral malaria or severe anaemia).1 2 To assess the role of these genes in mild disease, we conducted a genetic analysis of sibling pairs concordant for this phenotype. The affected sib-pair method compares the observed and expected distribution of parental alleles at marker loci inherited identical by descent (ibd). At any locus, a pair of siblings may share 0, 1, or 2 alleles in the ratio 25%:50%:25%, by random segregation. If a locus is genetically linked to disease, affected siblings will share a higher number of alleles identical by descent at that locus than expected.

Subjects, methods, and results

We recruited 217 dizygous pairs of Gambian twins (mean age 5.3 years).3 Twins living together are important for such a study, for they are of the same age and share a common environment. They were monitored weekly during the 1991 rainy season for development of fever and P falciparum infection. Clinical malaria was defined as fever (axillary temperature ≥37.5°C) plus P falciparum asexual parasitaemia ≥5000/μl. Surveillance continued through the next two rainy seasons, producing a total of 40 pairs of twins who were concordant for clinical malaria; none had severe disease.

Major histocompatibility complex class II typing of these children and their parents was by TaqI restriction fragment length polymorphism analysis; the cDNA probes used for hybridisation were specific for DRB and for DQB. Additional typing was by fluorescence labelled microsatellite markers that map to the major histocompatibility complex region (d6s291, d6s273, tnfa, d6s276). A maximum of four indi- vidual alleles were labelled within each family. Analysis was by GAS (Genetic Analysis System, version 1.6; Alan Young, Oxford) and the MAPMAKER/SIBS program, which uses a maximum likelihood method that takes into account partially informative matings to infer sharing of alleles identical by descent at the unknown loci and thereby to compute, at each location, a maximum lod score (log 10 of the likelihood of estimated sharing divided by the likelihood of no excess sharing).4

Four families had moved away from the study area before the end of the study, and parents in four families had inadequate samples; in nine families one parental sample was not available, and in one family a parent was homozygous at all of the loci tested. The 22 fully informative families were analysed by the gas program (table 1).

Table 1

Distribution of shared alleles mapping to the major histocompatibility complex by affected sibling pairs concordant for malaria in 22 families with complete information (based on the single most informative marker for each family)

View this table:

The MAPMAKER/SIBS program estimated the maximum likelihood values of the allele sharing proportions for each marker for 32 families (the 22 above plus the 10 families with incomplete data) to estimate the maximum lod score, which was 2.77 (P<0.001) at the tnfa locus. For comparison, sharing of major histocompatibility complex alleles was not increased among 13 pairs of dizygous twins who were discordant for clinical disease (only one member of each pair having developed clinical malaria during the three year observation period) (data not shown).


The significantly non-random sharing of alleles indicates a large effect of the major histocompatibility complex on risk of uncomplicated malaria. The linkage approach may be a more powerful means of assessing the overall influence of the complex in infectious diseases, for previous studies of uncomplicated malaria have detected no association with individual alleles.5 Three genes of the major histocompatibility complex are already known to affect the outcome of malaria infections; others may remain to be identified.


We thank Alan Young for providing the GAS program and Martin Farrall for carrying out the analysis with the MAPMAKER/SIBS program.

Funding: UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases; Wellcome Trust; Rockefeller Foundation; Medical Research Council.

Conflict of interest: None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract