DNA Mixture Deconvolution: A Four-Strategy Framework from Physical Separation to Database Searching.

Area of Science:

• Forensic Genetics
• Molecular Biology
• Computational Biology

Background:

• DNA mixture interpretation is a complex forensic challenge.
• Probabilistic genotyping (PG) advanced likelihood ratio (LR) evaluation but not contributor genotype reconstruction.
• Limited focus on systematic genotype reconstruction in database searches.

Purpose of the Study:

• Synthesize recent developments in DNA mixture deconvolution.
• Review strategies for improving contributor genotype reconstruction.
• Discuss the integration of deconvolution with database searching.

Main Methods:

• Physical and biological separation techniques (e.g., single-cell isolation).
• High-information genetic markers (microhaplotypes, MiniHaps, DIP-STRs).
• Continuous probabilistic algorithms (Hamiltonian Monte Carlo, variational inference, deep learning).

Main Results:

• Upstream separation reduces mixture complexity.
• Novel markers increase information content and degradation resistance.
• Advanced algorithms improve genotype inference stability and accuracy.
• Deconvolution and LR evaluation are distinct objectives requiring separate validation.

Conclusions:

• Convergence of molecular innovation and algorithms transforms mixture interpretation into an investigative framework.
• Future progress requires standardized markers, deconvolution metrics, and scalable database infrastructures.
• Enhanced DNA deconvolution aids forensic investigations and database searching.