DNA残留在RNA提取过程中难以避免，并在测序过程中可被富集，影响RNA-seq结果的可靠性及临床应用。申请人前期研究发现，DNA污染普遍存在于RNA-seq数据中，且具有强烈的基因组区域偏好性；但目前尚无相关的系统研究，更无消除其影响的计算方法。本项目拟设计已知浓度DNA污染的RNA标准样品系列，产生相应的RNA-seq数据，从而：1.建立从RNA-seq数据估算DNA污染程度的校正模型；2.在全基因组区域鉴定并量化DNA污染的分布规律及序列偏好性特征；3.建立消除DNA污染的计算方法，并采用SEQC等标准数据集评估和优化方法；4.用此新算法重新分析ENCODE等数据集，比较分析结果的异同并进行实验验证，评估算法的有效性。本项目将系统地阐明DNA污染对RNA-seq结果准确度的影响，所建立的消除DNA污染影响的新算法可望广泛用于处理未来及既往RNA-seq数据，提高分析结果可靠性。

Residual genomic DNA (gDNA) is present in RNA isolation process and could be enriched during some RNA-seq protocols, impacting the reliability of RNA-seq results. We observed that gDNA contamination in RNA-seq data is ubiquitous and there exist sequence biases across the whole genome. However, there is no systematic study on the characteristics of such biases, and no effective experimental or computational methods are available to characterize and remove such negative impact from gDNA contamination. In this proposed study, RNA-seq data from standard RNA samples with gDNA of known concentrations will be generated in order to: 1. establish calibration models for estimating the level of gDNA contamination from RNA-seq data; 2. characterize and quantify the profiles and sequence biases of gDNA contamination at the genome scale; 3. develop a data analysis pipeline for minimizing the negative impact of gDNA contamination in RNA-seq data; 4. optimize and assess the performance of the data analysis pipeline using the huge RNA-seq data sets from the SEQC project, and apply the data analysis method to well-known data sets such as ENCODE to evaluate the degree of gDNA contamination and to validate a subset of analysis results. The proposed method may become suitable for analyzing both historical and future RNA-seq data and is expected to significantly improve the reliability of RNA-seq measurements.