A hybrid approach to increase the informedness of CE-based data using locus-specific thresholding and machine...

Area of Science:

• Forensic Science
• Genetics
• Bioinformatics

Background:

• Interpreting genetic profiles requires distinguishing true allelic information from noise.
• Static analytical thresholds in capillary electrophoresis data analysis can lead to loss of valuable allelic information.
• Traditional methods struggle to adapt to variability in instrumentation and sample conditions.

Purpose of the Study:

• To develop a robust and adaptive method for genetic profile interpretation.
• To minimize false positives (artifact detection) and false negatives (allele dropout) in genetic data.
• To improve allele detection accuracy over traditional static thresholds.

Main Methods:

• Implemented a dynamic, locus- and sample-specific analytical threshold.
• Utilized a machine learning-derived probabilistic artifact detection model.
• Compared the adaptive method against static thresholds using capillary electrophoresis data.

Main Results:

• Achieved an allele detection accuracy of 97.2%, an 11.4% improvement over the lowest static threshold.
• Reported a low incidence of incorrectly identified artifacts at 0.79%.
• Demonstrated superior retention of allelic information content compared to static thresholds.

Conclusions:

• The adaptive method significantly outperforms static thresholds in genetic data analysis.
• This approach offers a more accurate and comprehensive interpretation of genetic profiles.
• The dynamic and machine learning-based system provides a cost-effective solution for reducing data loss.