A Tool for Predicting Haemoglobin Variants - an Aid to DNA Diagnostics.

The hemoglobin (Hb) variants represent a well-documented array (>800 in number) of amino acid substitutions, hybrid Hbs and Hbs with extended/shortened amino acid sequences. The majority of these structural alterations have little or no impact on the well being of the individual. However, some can result in severe disease when inherited in a homozygous or compound heterozygous state (e.g. sickle cell disease), while others can give a thalassaemic picture if the globin variant is highly unstable. Kings’s College Hospital (KCH), which is a referral centre for Hb variants, uses automated HPLC (Biorad Variant II) as the first line screening test, followed by further electrophoretic separations (isoelectric focusing (IEF), cellulose acetate membrane (CAM) and acid gel electrophoresis (AGE)) if a variant is present. Definitive diagnosis is then achieved by either mass spectrometry or DNA analysis of the suspected globin gene. The former method is quick and uses small quantities of blood but the equipment is expensive and requires a high degree of expertise, making it unavailable to many institutions. PCR based DNA diagnostics is routine in many service pathology laboratories but it is hampered by the cost of consumables and the time it takes. To minimise these KCH has developed a database that statistically evaluates the phenotypic separations and predicts which Hb variant is present based on past data. A single entry from any one of the tests (HPLC, IEF, CAM, AGE or mass difference) is enough to generate a prediction. Using this prediction only a small region of the putative globin gene will need sequencing or analysis by restriction digest, thereby reducing costs and turnaround times. The database contains the phenotypic separation data of >700 samples collected over six years. Variants were definitively diagnosed using electrospray ionisation mass spectrometry, complemented in some cases by DNA analysis. KCH will continue to enter variant data to increase the data set and improve the tool’s accuracy and aims to make the prediction tool available online so that other laboratories may obtain predictions and add their data. Laboratories will then be able to retrieve predictions based on their own past data or the entire database.