Abstract —We have analyzed the gene expression dynamics in floral transition in the CDC Frontier chickpea cultivar. We provide a model, in several versions, to predict the expression dynamics of five flowering genes, taking the expression of their regulators as an input. The models were trained using the random forest method on the previously published expression data for ten flowering genes under the short- and long-day growing conditions. The resulting models correctly predict the dynamics of the average expression levels under long days. We show that the models for CDC Frontier mainly reproduce the regulatory interactions between the key genes described for the Arabidopsis thaliana model plant. Based on the analysis, we hypothesize that the short-day data and the long-day data contain qualitatively different information, which may be due to different regulatory modules that function in different conditions. For the regulators of the flower meristem identity genes AP1 and LFY , our models predict FTa3 as the main activator and TFL1c as the main repressor under long days.

中文翻译：

---

CDC Frontier 鹰嘴豆品种开花基因表达的机器学习分析

摘要——我们分析了 CDC Frontier 鹰嘴豆品种花期转变中的基因表达动态。我们提供了多个版本的模型，以预测五个开花基因的表达动态，并将其调节因子的表达作为输入。使用随机森林方法对先前公布的十个开花基因在短日照和长日照生长条件下的表达数据进行训练。由此产生的模型正确预测了长日下平均表达水平的动态。我们表明，CDC Frontier 模型主要再现了拟南芥模式植物描述的关键基因之间的调控相互作用。基于分析，我们假设短日数据和长日数据包含质性不同的信息，这可能是由于在不同条件下运行的不同监管模块。对于花分生组织特征基因 AP1 和 LFY 的调节因子，我们的模型预测 FTa3 是长日照下的主要激活因子，TFL1c 是主要的抑制因子。