SorTR (automated reference DNA analysis and rework selection)

ESR interprets DNA profiles from graphs of different coloured peaks (called Electropherograms (EPGs)) that indicate the presence of pieces of DNA of different sizes. The process of DNA profiling causes artefacts so resulting EPGs consist of a baseline signal with a number of peaks which may be artefacts or true DNA fragments. Analysis is required to remove all artefacts based on a set of rules leaving only allelic peaks. However, the interpretation of EPGs can be difficult, consuming resources and time. In addition, there can be a number of reasons why suitable profiling results are not achieved, and therefore an analyst needs to make the decision on the appropriate rework for each failed sample.

SorTR is a prototype workflow using two machine learning models that automatically interpret reference DNA profiles to assign the alleles present and, where required, recommends appropriate rework options. Utilising this prototype system would reduce licence costs and create a minimum time saving of approximately 27 work days in a year.

Janet Stacey (presenting author) – Digital Sciences Engineer (ESR)

Anna Lemalu - Senior Scientist – Forensic Biology (ESR) Maria van der Salm – Scientist – Forensic Biology (ESR)