The Promise of Explainable Deep Learning for Omics Data Analysis: Adding New Discovery Tools to AI

Abstract

Deep learning has already revolutionised the way we process a wide range of data, in many areas of our daily life. The ability to learn abstractions and relationships from heterogeneous data, has provided impressively accurate prediction and classification tools to handle increasingly big datasets. This has a significant impact on the growing wealth of omics datasets, with the unprecedented opportunity for a better understanding of the complexity of living organisms. While this revolution is transforming the way we analyse these data, explainable deep learning is emerging as an additional tool with the potential to change the way we interpret biological data. Explainability addresses critical issues such as transparency, so important when computational tools are introduced especially in clinical environments. Moreover, it empowers artificial intelligence with the capability to provide new insights in the input data, thus adding an element of discovery to these already powerful resources. In this review we provide an overview of the transformative effects explainable deep learning is having on multiple sectors, ranging from genome engineering and genomics, from radiomics to drug design and clinical trials. We offer a perspective to life scientists, to better understand the potential of these tools, and a motivation to implement them in their research, by suggesting learning resources they can use to move their first steps in this field.

Accurate determination of the meaning and implications of traditional Chinese medicine using panoramic spatiotemporal life atlas

“…XGBoost 决策规则与决策树相同，支持回归和分类，属于一种采用树结构设计的回归 树，与传统的梯度提升决策树(Gradient Boosting Decision Tree, GBDT)相比，优化了计算 速度， 并且在目标函数上加入了惩罚项大大增强了模型的泛化能力， 已广泛应用于分子诊断、 建模等数据挖掘领域 [46] ，对于样本的评估与预测具有较高的灵敏度、精确度和召回率。…”

“…等算法)、聚类(K-means、 graph-based、louvain 等算法)、数据归一化标准化(SCT 算法)、非监督聚类(Markvariogram 算法)、多模式交叉分析、相关性分析、反卷积(SPOTlight)、有监督学习(RCTD、spotlight 算法)、神经网络模型、空间形态学降维聚类(stSME)、空间收缩质心聚类(计算像素点 所有对应的近邻距离基础上做 K-means 聚类)、伪时空轨迹推断算法(CCI、PST)、空间 近邻分析、富集分析、线性回归(Pearson)、判别分析(OPLS-DA)等[43][44][45][46][47][48][49] ，下面对相对复 杂的几种方法展开详细介绍。 A c c e p t e d 图 5 人工智能分类[44][45][46][47][48][49] …”

Interpretable deep learning framework for understanding molecular changes in human brains with Alzheimer’s disease: implications for microglia activation and sex differences

“…b. For each feature ‫ݔ‬ , computed the mean SHAP value for CN samples as equation (3), where ܰ ே represents the total number of CN samples in the dataset.…”

“…One of the challenges faced by the application of AI approaches to multi-omic data is lack of interpretability. Complex ML models, like Deep Neural Networks (DNNs), although with unparalleled predictive power, are often considered as “black box” models as their decision- making processes are not easily inspected by human investigators[3]. Existing literature has reported the use of different AI frameworks to uncover deep interrelationships between gene expression and AD neuropathologies[4, 5].…”