Related Experiment Videos
Optimal structure for automatic processing of DNA sequences
S W Davies1, M Eizenman, S Pasupathy
1Bell Laboratories, Holmdel, NJ 07733-3030, USA.
IEEE Transactions on Bio-Medical Engineering
|September 24, 1999
Summary
This study introduces a statistical model and a maximum-likelihood (ML) algorithm for DNA sequencing, improving base sequence recovery from electrophoresis data. The DNA-ML algorithm enhances accuracy by using Kalman prediction and advanced statistical processing.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- Automatic DeoxyriboNucleic Acid (DNA) sequencing accuracy relies on statistical analysis of electrophoresis time series data.
- Current DNA sequencing algorithms are largely heuristic, with limited statistical information utilization.
Purpose of the Study:
- To develop a formal statistical model for DNA electrophoresis time series.
- To construct an optimal maximum-likelihood (ML) processor based on this model for improved DNA sequencing.
Main Methods:
- Developed a formal statistical model for DNA time series.
- Constructed an optimal maximum-likelihood (ML) processor, termed DNA-ML.
- Incorporated Kalman prediction, peak parameter estimation, whitened waveform comparison, and M-algorithm for multiple hypothesis processing.
Main Results:
- Examined DNA-ML algorithm properties using simulated and real data.
- Identified critical model parameters influencing error mechanisms like insertions and deletions.
- Demonstrated the effectiveness of the statistical model and DNA-ML algorithm.
Conclusions:
- The developed statistical model and DNA-ML algorithm offer a robust framework for DNA sequencing.
- This approach provides a foundation for future advancements in DNA sequencing techniques and error reduction.