Predicting the clinical outcome of infectious and non-infectious disease patients using whole blood transcriptomics

患有相同疾病（传染性或非传染性）的个体之间的临床差异很大，这是绝大多数疾病的医学现实。这些临床表现可以从无症状到致命，其间有一系列临床并发症。例如，受疟疾感染的病人一般分为无并发症感染和严重感染两类。该病的严重和致命形式几乎完全由恶性疟原虫引起，其特点是存在一种或多种临床并发症。这些疾病包括：脑性疟疾、严重贫血、肾衰竭、低血糖、乳酸酸中毒、肺水肿、呼吸窘迫综合征和肝功能障碍。因此，这种疾病的严重程度因人而异，但70%的疟疾死亡发生在5岁以下儿童中。因此，转录组学，主要基于差异表达的方法已被用于研究这种疾病和其他疾病，以确定转录组预后生物标志物。这导致确定了一些途径，包括toll样受体途径，以及与疟疾严重程度相关的基因，但从简单感染到严重感染的进展仍然知之甚少。因此，我们无法根据我们在临床表现时所能测量到的情况来预测患者因恶性疟原虫感染而可能出现的临床并发症。因此，可能需要新的方法来模拟病人疾病的自然历史。一种可以用于此目的的方法是命运映射。命运图谱，在单细胞水平上进行时也被称为细胞谱系追踪，它使用细胞的转录组谱来追踪其发育过程，通常是分化。最近，这种方法与一种新的方法——RNA速度（RNA velocity）相结合，将方向性置于图谱中每个细胞的转录组上（velocyto算法）。我在这项研究中使用的就是这种方法。

High clinical variability between individuals with the same (infectious or non-infectious) disease is the medical reality of the vast majority of diseases. These clinical manifestations can range from asymptomatic to lethal with a range of clinical complications in between. For example, malaria infected patients are generally divided into those with either uncomplicated or severe infections. The severe and deadlier form of the disease is almost exclusively caused by Plasmodium falciparum and is characterised by the presence of one or more clinical complications. These can include: cerebral malaria, severe anaemia, kidney failure, hypoglycaemia, lactic acidosis, pulmonary edema, respiratory distress syndrome and liver dysfunction. Consequently, there is high variability in the severity of this disease between individuals, but >70% of malarial deaths occur in children under 5. Hence, transcriptomic, primarily differential expression-based, methods have been utilised in the study of this and other diseases to identify transcriptomic prognostic biomarkers. This has resulted in the identification of a few pathways, including toll-like receptor pathways, and genes that are associated with malarial severity but the progression from an uncomplicated to severe infection remains poorly understood. Hence, we cannot predict the clinical complications a patient is likely to develop, as the result of a P. falciparum infection, based on what we can measure at clinical presentation. Thus, new approaches may be required to model the natural history of a patient's disease. One such approach that could be adapted for this purpose is fate mapping. Fate-mapping, which is also called cell lineage tracing when performed at the single cell level, uses a cell's transcriptomic profile to trace its development through a process, commonly differentiation. This has recently been combined with a new method, RNA velocity, to place a directionality onto the transcriptome of each cell within the map (the velocyto algorithm). It is this method I am using in this study.