P162 Use of an automated chimerism calculator to increase speed, precision, and statistical analysis of results

Aim Chimerism testing is a valuable tool for determining the proportion of donor vs. recipient DNA in a sample. The commonly used polymerase chain reaction amplifications of short tandem repeats assay (Promega, Madison, WI) produces results for 21 loci, but only 9 loci are used in manual calculations. We sought to improve the tedious and time consuming process of manually calculating a patient’s percent chimerism by developing an automated chimerism calculator for rapid and accurate chimerism calculations based off of all 21 loci. Methods The calculator was developed using Microsoft Excel Macros and is able to report the percent donor chimerism +/− the margin of error. To determine whether precision is improved by using 21 loci, the results of a 21 loci analysis and a 9 loci analysis were compared in 111 patient samples. Additionally, each of the 21 loci was evaluated for its average deviation from the mean. Finally, 298 patient samples were analyzed by the program in order to determine if precision is equal at all chimerism percentages. Results When the same loci are analyzed, the chimerism calculator results are identical to the manual calculations, but completed in 10% of the time. The analysis of patient samples using 21 and 9 loci yielded, on average, chimerism percentages within 2.0% of each other. The average margin of error using 9 loci for analysis is 2.41% compared to 0.8% using all 21 loci. Among the 21 loci, the average deviation from the mean was found to range from 8.9% and 5.9%. Finally, we found that the average margin of error for patients between 90 and 100 percent donor chimerism is 0.19% while patients between 30 and 40 percent donor chimerism have 10.1% margin of error. Conclusions The chimerism calculator allows for routine use of all 21 STR loci with significantly reduced analysis time, 30% less margin of error, and an informative statistical analysis. In various laboratory settings, similar methods of computer assisted chimerism analysis can be easily implemented.