Modelling the link between replication stress, chromosomal instabilityand aneuploidy

染色体不稳定(CIN)是一种常见于癌症和各种遗传性疾病的疾病，其定义为染色体结构或整体变异率增加。导致染色体错误分离的有丝分裂缺陷通常被认为是整个染色体CIN(W-CIN)的根本原因，而结构性CIN(S-CIN)则被认为是复制应激(RS)所致。我们发现包括GINS1和CDC45在内的复制基因在W-CIN肿瘤中过度表达，这一普遍观点受到了挑战。事实上，与这一观察结果一致的机制已经与Holger bastians(FOR2800，SP2)合作确定：由这些基因或轻度RS触发的复制起点激发增加足以诱发W-CIN。RS如何触发起源激发的增加以及复制机制的这些干扰如何诱导CIN的细节目前尚不清楚。我们将开发和校准一个关于人类细胞中DNA复制和起始激发的机械数学模型。我们将使用来自DNA梳理实验的数据，提供有关分叉速度和原点间距离的信息，以约束这个模型的参数，以及我们自己的处理结构模型错误的数学技术，以建议进一步的数据采集。通过在现象学上将该模型与有丝分裂缺陷直接联系起来，我们将能够模拟复制机制的各种干预和扰动(例如GINS1过表达)对起源发火率、有丝分裂染色体分离和CIN的影响。然后，该模型的第二个版本将包括ATM路径的模型。我们的目标是通过计算机模拟来研究ATM系统如何将起源激发和复制动力学的变化转移到有丝分裂缺陷和CIN。然后，这些模型预测将由FOR2800的成员进行实验。由于非整倍体和CIN通常对细胞健康有不利影响，我们的目标是了解耐受CIN所需的纵向变化，以确定CIN在癌症中的作用。在与Zuzana Storchova(FOR2800，SP8)的合作中，我们将修改我们最近开发的MFmap机器学习框架，以集成高维基因组数据，并研究在适应非整倍体和CIN期间细胞状态的变化。MFmap框架将允许我们预测癌症亚型标签，并计算适应期间不同时间点样本的低维潜在表示。通过研究这些变化，并通过将潜在空间的变化与增殖特征相关联，我们的目标是识别和分类CIN适应过程中的关键事件。

Chromosomal instability (CIN), a condition frequently observed in cancer and in variou inheritable diseases, is defined as an increased rate of structural or whole chromosome alterations. Mitotic defects triggering chromosome missegregation are often considered as theroot cause of whole chromosome CIN (W-CIN), whereas structural CIN (S-CIN) is attributed to replication stress (RS). This prevailing view has been challenged by our finding that replication genes including GINS1 and CDC45 are overexpressed in W-CIN tumours. Indeed, a mechanism consistent with this observation has been identified in collaboration with Holger Bastians (FOR2800, SP2): increased replication origin firing triggered by these genes or by mild RS is sufficient to induce W-CIN. The details of how the increased origin firing is triggered by RS and how these disturbances of the replication machinery induce CIN are currently unexplored. We will develop and calibrate a mechanistic mathematical model for DNA replication and origin firing in human cells. We will use data from DNA combing experiments providing information on the fork speed and inter-origin distance to constrain the parameters of this model and our own mathematical techniques for handling structural model errors to suggest further data acquisition. By directly linking this model phenomenologically to mitotic defects, we will be able to simulate the effect of various interventions and perturbations of thereplication machinery (for instance GINS1 overexpression) on origin firing rate, mitotic chromosome segregation and CIN. A second version of the model will then include a model of the ATM pathway. We aim to study by computer simulations, how the ATM system transfer changes in origin firing and replication dynamics to mitotic defects and CIN. These model predictions will then be addressed experimentally by members of the FOR2800. Since aneuploidy and CIN have typically detrimental effects on cellular fitness, we aim to understand the longitudinal changes required to tolerate CIN to define the role of CIN in cancer. In collaboration with Zuzana Storchova (FOR2800, SP8), we will modify our recently developed MFmap machine learning framework to integrate high dimensional genomic data and to studychanges of the cell state during the adaptation to aneuploidy and CIN. The MFmap framework will allow us to predict cancer subtype labels and to compute low dimensional latent representations for samples at different time points during adaptation. By studying thesechanges and by correlating changes in the latent space to proliferation signatures we aim to identify and categorise key events along the adaptation process to CIN.