Methodology

MalHaploFreq: A computer programme for estimating malaria haplotype frequencies from blood samples

Ian M Hastings^1* and Thomas A Smith²

• * Corresponding author: Ian M Hastings hastings@liverpool.ac.uk

Author Affiliations

¹ Liverpool School of Tropical Medicine, Pembroke Place, Liverpool, L3 5QA, UK

² Swiss Tropical Institute, Socinstrasse 57, CH-4002 Basel, Switzerland

For all author emails, please log on.

Malaria Journal 2008, 7:130 doi:10.1186/1475-2875-7-130

Published: 15 July 2008

Abstract

Background

Molecular markers, particularly those associated with drug resistance, are important surveillance tools that can inform policy choice. People infected with falciparum malaria often contain several genetically-distinct clones of the parasite; genotyping the patients' blood reveals whether or not the marker is present (i.e. its prevalence), but does not reveal its frequency. For example a person with four malaria clones may contain both mutant and wildtype forms of a marker but it is not possible to distinguish the relative frequencies of the mutant and wildtypes i.e. 1:3, 2:2 or 3:1.

Methods

An appropriate method for obtaining frequencies from prevalence data is by Maximum Likelihood analysis. A computer programme has been developed that allows the frequency of markers, and haplotypes defined by up to three codons, to be estimated from blood phenotype data.

Results

The programme has been fully documented [see Additional File 1] and provided with a user-friendly interface suitable for large scale analyses. It returns accurate frequencies and 95% confidence intervals from simulated dataset sets and has been extensively tested on field data sets.

Conclusion

The programme is included [see Additional File 2] and/or may be freely downloaded from [1]. It can then be used to extract molecular marker and haplotype frequencies from their prevalence in human blood samples. This should enhance the use of frequency data to inform antimalarial drug policy choice.