Efficient collection, management and dissemination of information are essential to intervene effectively against neglected tropical diseases and to monitor and evaluate strategies and actions towards their control. However, the neglect suffered by many diseases that affect the world's poorest people manifests itself in many guises, not least in the paucity of reliable information on their incidence and geographic distribution [1, 2]. Evidence on the number and location of infections is badly needed to bring these neglected diseases out of the shadow  and, most crucially, to tackle them efficiently [4, 5].
Disease maps are required for planning, managing and monitoring interventions across the whole spectrum of neglected tropical diseases. For conditions such as helminthiases and trachoma, against which safe and affordable drugs are available, baseline maps of endemicity are used to plan mass drug administration (MDA) campaigns [6–8]. As campaigns progress, spatially-explicit data enable implementation to be monitored and ultimately contribute to assessing the degree of success of the initiatives . Geo-spatial information also plays a role in the post-MDA phase when surveillance aims at providing early-warning of the possible disease resurgence or reintroduction in freed areas.
A different group of diseases such as human African trypanosomiasis (HAT), Chagas disease and leishmaniasis, are characterized by the limited availability of safe and cost-effective control tools. This group includes infections of complex epidemiology, which, if untreated, can result in very high death rates. Strategies to control them hinge on early diagnosis and treatment, as well as on vector control. Accurate, spatially-explicit information is essential to target, monitor and evaluate interventions against these challenging and deadly diseases [10–12]
Geo-referenced data are also crucial when control strategies are successful and elimination or eradication of a disease can be contemplated. This is exemplified by the "Dracunculiasis Eradication Program", wherein village-level information on disease occurrence and programme implementation is routinely collected to monitor progress towards the programme's ultimate goal .
Disease distribution maps are also required if accurate estimates of the population at risk and of the burden of diseases are to be made. The issue of the most appropriate metrics to estimate the burden is subject of debate [14, 15], but there is increasing recognition that the burden of neglected tropical diseases still needs to be fully appreciated  and that more detailed and reliable data on their incidence and impacts need to be collected [17–19].
The present paper focuses on HAT, also known as sleeping sickness, whose etiological agents are tsetse-transmitted protozoa of the species Trypanosoma brucei. Two sub-species are responsible for syndromes of markedly different epidemiology and geographical range. T. b. gambiense causes the chronic, anthroponotic form found in Western and Central Africa. Infection with T. b. rhodesiense results in the acute, zoonotic form of Eastern and Southern Africa. Both forms are considered fatal if untreated. No vaccine against HAT is available, and the toxicity of existing drugs precludes the adoption of control strategies based on preventive chemotherapy. As a result, the keystones of interventions against sleeping sickness are active and passive case-finding followed by treatment, vector control and animal reservoir management.
Following World Health Assembly resolutions 50.36 in 1997  and 56.7 in 2003 , the World Health Organization (WHO) committed itself to supporting HAT-endemic countries in their efforts to remove the disease as a public health problem. Mapping disease distribution in time and space has a pivotal role to play if this objective is to be met.
Partly because of the focal nature of sleeping sickness geography, HAT control has always been intimately entwined with disease mapping . For example, as early as 1903, detailed maps of HAT distribution were published in Uganda  and foci were mapped in Western and Central Africa in the period 1908-1910 [24, 25]. In the last decade, mapping and geo-positioning have become much more accessible with the widespread diffusion of Geographic Information Systems (GIS) and satellite-aided positioning systems (e.g. the Global Positioning System (GPS)). These technical tools are at the basis of the systematic approach to sleeping sickness mapping embodied in the Atlas of HAT, the first attempt ever to geo-reference at the village level all sleeping sickness cases reported from affected countries . Underpinning this initiative are the commitment of WHO and the Food and Agriculture Organization of the United Nations (FAO), the institutional umbrella provided by the Programme against African Trypanosomosis (PAAT) and the crucial input of National Sleeping Sickness Control Programmes (NSSCPs), Non-Governmental Organizations (NGOs) and Research Institutes.
This paper presents the process of building the Atlas of HAT for the period 2000-2009. The distribution of HAT is presented for 23 out of 25 sub-Saharan countries having reported on the status of sleeping sickness during the last ten years: Benin, Burkina Faso, Cameroon, Central African Republic, Chad, Congo, Côte d'Ivoire, Equatorial Guinea, Gabon, Ghana, Guinea, Kenya, Malawi, Mali, Mozambique, Nigeria, Rwanda, Sudan, Togo, Uganda, United Republic of Tanzania, Zambia and Zimbabwe. Work is in progress to process data from the two remaining countries having reported on HAT in the study period: Angola and the Democratic Republic of the Congo (DRC).
The presented maps are a fundamental milestone in the process of building the continental Atlas of HAT. The unprecedented accuracy and comprehensiveness of the Atlas are to make it an essential tool for disease control, research and advocacy.