Unravelling reference bias in ancient DNA datasets

Area of Science:

• Ancient Genomics
• Bioinformatics
• Computational Biology

Background:

• Ancient genome characterization relies on accurate sequencing read alignment.
• Reference bias and spurious mappings are significant challenges in ancient DNA (aDNA) analysis.
• These issues are influenced by read length, mapping tools, and reference genomes.

Purpose of the Study:

• To investigate the impact of mapping tools, quality thresholds, and reference genomes on mismatch rates across varying read lengths in palaeogenomic datasets.
• To introduce AMBER, a bioinformatics tool for assessing aDNA mapping quality and identifying biases.

Main Methods:

• Analysis of empirical and simulated palaeogenomic datasets.
• Development and application of the AMBER bioinformatics tool.
• Assessment of mapping bias, mismatch rates, cytosine deamination, fragment length, and coverage.

Main Results:

• Mapping algorithms and quality thresholds predictably influence reference bias and spurious alignment rates based on read length.
• AMBER provides rapid assessment of aDNA mapping quality, including mismatch rates per read length and deamination profiles.
• Optimized mapping parameters for specific read lengths are key to mitigating reference bias and spurious mappings.

Conclusions:

• The choice of mapping parameters significantly impacts the accuracy of ancient genome analysis.
• AMBER is a valuable tool for quality control and optimization of aDNA mapping pipelines.
• Addressing reference bias and spurious mappings through parameter optimization is essential for reliable palaeogenomic studies.