My scientific career has mainly consisted of crossing roads within the locale of University College London. After studying Biochemistry there as an undergraduate, I crossed Gower Street and obtained a PhD on Mammalian Embryonic Haemoglobins in the Department of Clinical Haematology, University College Hospital Medical School, focussing on the human embryonic haemoglobins Hbs Gower 1 and Gower 2. Crossing another road, I then spent nearly seven years running a NHS diagnostic haematology laboratory in University College Hospital, screening for haemoglobinopathies and red cell enzymopathies and investigating haemolytic anaemias, although one year of this time was spent rather further afield at Groote Schuur Hospital in Cape Town. My foray into the world of molecular biology followed, with a move back into the Medical School to join the Haemopoiesis research group headed by Professor David Linch, now part of UCL Cancer Institute. This involved a shift in emphasis towards investigation of myeloid disorders. My early studies, predating the wealth of information now available on specific molecular abnormalities, applied clonal analysis using X-chromosome inactivation patterns in females to the biology of haematological neoplasms. They raised a number of issues for interpretation of the results that evolved into a broader understanding of haemopoietic stem cell biology, for example, the change in stem cell usage with age. They demonstrated that the myeloproliferative neoplasm ET is heterogeneous, with patients having either clonal or polyclonal myelopoiesis, and more recently we have associated this biological difference with the underlying calreticulin or JAK2 gene mutation. However, my major area of study has been the investigation of molecular mutations or abnormalities in young adult patients with AML and their impact on prognostic significance, risk stratification and potential therapeutic management. This work has been facilitated by our access to samples from patients entered into the national MRC/NCRI AML trials dating back to 1988, thereby allowing robust statistical analysis of large cohorts of homogenously treated patients with extensive clinical follow-up data, and these studies have highlighted a number of key factors that reflect the heterogeneity of the mutational profile in AML and the challenges for applying this information to risk stratification.