Comprehensive overview of various haplotype block inference methods

During last years there has been a breakthrough in genetics and biotechnology. Present technology allows large scale association studies between genotypes and diseases. The most powerful methodology of association studies is the analyse of single nucleotide polymorphisms (SNPs). The estimated number of SNPs in the human genome is 10 million. Therefore, SNP markers provide practical fine-grain map of human genome. Theoretical models and practical experiments suggest that a single strand of DNA is composed from haplotype blocks—sequences of fixed DNA strands. In other words, few suitably chosen SNPs can reveal most of DNA strand and reduce the number of required SNPs for association studies. The following review gives a introduction to basic concepts and models. We analyse pros and cons of various definitions of haplotype blocks, point out advantages and limitations of inference methods. We cover all three main approaches: methods based on marker pairs, combinatorial methods and models with minimal-description length. We also briefly discuss the optimal marker set (OMS) problem that is a common for all methodologies. Good solution can provide high quality of the end results and lower the measurement costs. However, the complete treatment of OMS is out of our scope.