Deconvolution of Nascent Sequencing Data Using Transcriptional Regulatory Elements.

Area of Science:

• Genomics and Molecular Biology
• Bioinformatics and Computational Biology

Background:

• Microdissection of heterogeneous tissues is crucial for biological and biomedical research.
• Supervised deconvolution for mixed sequencing data was previously limited to RNA-seq and ATAC-seq.

Purpose of the Study:

• To develop supervised deconvolution methods for nascent sequencing data (GRO-seq, PRO-seq).
• To identify informative features and assess deconvolution method performance for active transcription analysis.

Main Methods:

• Applied supervised deconvolution to nascent sequencing data for the first time.
• Developed a novel filtering method for promoter and enhancer regions.
• Utilized in-silico mixtures and RNA-seq best practices for validation.

Main Results:

• Enhancer RNAs were identified as highly informative features for deconvolution.
• Simpler deconvolution methods generally outperformed complex ones for nascent sequencing data.
• Undifferentiated cell types significantly confounded deconvolution results due to open chromatin.

Conclusions:

• Supervised deconvolution of nascent sequencing data is generally feasible.
• Future approaches may require integration with other sequencing protocols to address challenges with undifferentiated cell types.