DAISM-DNNXMBD: Highly accurate cell type proportion estimation with in silico data augmentation and deep neural networks

Yating Lin; Haojun Li; Xu Xiao; Lei Zhang; Kejia Wang; Jingbo Zhao; Minshu Wang; Frank Zheng; Minwei Zhang; Wenxian Yang; Jiahuai Han; Rongshan Yu

doi:10.1016/j.patter.2022.100440

DAISM-DNN^XMBD: Highly accurate cell type proportion estimation with in silico data augmentation and deep neural networks

Patterns (N Y). 2022 Feb 3;3(3):100440. doi: 10.1016/j.patter.2022.100440. eCollection 2022 Mar 11.

Authors

Yating Lin¹, Haojun Li¹, Xu Xiao^{1

2}, Lei Zhang³, Kejia Wang⁴, Jingbo Zhao⁵, Minshu Wang^{2

4}, Frank Zheng⁵, Minwei Zhang⁶, Wenxian Yang⁷, Jiahuai Han^{2

3

8}, Rongshan Yu^{1

2

7}

Affiliations

¹ School of Informatics, Xiamen University, Xiamen 361005, China.
² National Institute for Data Science in Health and Medicine, Xiamen University, Xiamen 361005, China.
³ School of Life Science, Xiamen University, Xiamen 361102, China.
⁴ School of Medicine, Xiamen University, Xiamen 361102, China.
⁵ Amoy Diagnostics, Xiamen, 361000, China.
⁶ Department of Critical Care Medicine, The First Affiliated Hospital of Xiamen University, Xiamen 361003, China.
⁷ Aginome Scientific, Xiamen, 361005, China.
⁸ Research Unit of Cellular Stress of CAMS, Cancer Research Center of Xiamen University, School of Medicine, Xiamen University, Xiamen 361102, China.

Abstract

Understanding the immune cell abundance of cancer and other disease-related tissues has an important role in guiding disease treatments. Computational cell type proportion estimation methods have been previously developed to derive such information from bulk RNA sequencing data. Unfortunately, our results show that the performance of these methods can be seriously plagued by the mismatch between training data and real-world data. To tackle this issue, we propose the DAISM-DNN^XMBD (XMBD: Xiamen Big Data, a biomedical open software initiative in the National Institute for Data Science in Health and Medicine, Xiamen University, China.) (denoted as DAISM-DNN) pipeline that trains a deep neural network (DNN) with dataset-specific training data populated from a certain amount of calibrated samples using DAISM, a novel data augmentation method with an in silico mixing strategy. The evaluation results demonstrate that the DAISM-DNN pipeline outperforms other existing methods consistently and substantially for all the cell types under evaluation in real-world datasets.

Keywords: cell type proportion estimation; data augmentation; data simulation; deconvolution; deep learning.