Visualizations for genetic assignment analyses using the saddlepoint approximation method

Area of Science:

• Population Genetics
• Bioinformatics
• Computational Biology

Background:

• Genetic assignment methods are crucial for understanding population structure and gene flow.
• Existing software often lacks effective visualization tools, hindering biological interpretation.
• Handling individuals with missing genetic data remains a challenge in assignment analyses.

Purpose of the Study:

• To develop an advanced method for visualizing genetic assignment data.
• To enhance existing genetic assignment techniques by accommodating missing genotype data.
• To provide a biologically interpretable visualization of population structure and assignment power.

Main Methods:

• Characterized genetic profile distributions for candidate source populations.
• Calculated graph positions for individuals with missing data using estimated quantiles.
• Employed the saddlepoint method to approximate and invert the cumulative distribution function (CDF) for quantile calculations.
• Utilized leave-one-out procedures for visualizing assignment results.

Main Results:

• The proposed visualization method effectively positions individuals with missing genetic data within population distributions.
• Saddlepoint approximation enabled accurate quantile function calculation for enhanced visualization.
• Applied to simulated and real microsatellite data (Rattus rattus), the method revealed population structure features.
• Demonstrated superior interpretability compared to existing bar chart visualizations.

Conclusions:

• The novel visualization method significantly advances genetic assignment analysis, particularly for datasets with missing data.
• It offers a more biologically meaningful interpretation of population structure and assignment confidence.
• The method provides a powerful tool for assessing the discriminative power of genetic markers.