Quantitative analysis of gel electrophoresis data for automated genotyping applications (Abstract)

Richards, D.R., and Perlin, M.W.

Methods for automated genotyping of polynucleotide repeat markers have recently been developed (Perlin et al, 1994). These methods genotype microsatellite data by deconvolution analysis of a marker's PCR stutter bands, thereby mathematically eliminating the stutter artifact and inferring bands that correspond to just the marker's actual allele sizes present in the sample. The key requirements for successful application of these deconvolution methods are accurate estimation of (a) the size in base pairs (bps) and (b) the relative DNA concentration for every band on the gel image. Available software analysis programs for automated DNA sequencers do not meet these requirements.

We have developed genetic analysis software in the MatLab programming language that meets these quantitative requirements. Analysis starts with a computer file of a gel electrophoresis image. For task (a) size calibration, known molecular weight markers having sufficient lane density are used to construct a mapping from the image into lanes and bp sizes. For task (b) band quantitation, the observed band shape is modeled; this model is applied to genetic data bands at expected image locations, and is used to determine their relative DNA concentration. Program output is a text file suitable for deconvolution analysis and allele calling. Our image analysis software was developed and proved successful for genotyping applications using ABI/373 gel collection data files. MatLab runs on Macintosh, PC, and UNIX systems.

in American Journal of Human Genetics, 57(4 Supplement): A26, 1995.
Back to FAST-MAP papers .
For comments, please contact perlin@cs.cmu.edu