Objective. The ornithine transcarbamylase is a mitochondrial matrix homotrimeric enzyme, whose deficiency is the most common genetic defect of the urea cycle and an X-linked semidominant disorder. To understand its mutation pattern is very helpful for managing its clinical manifest and outcome. Methods. The amino-acid pair predictability is used to transfer the symbolized human ornithine transcarbamylase and its 117 missense point mutants to scalar data and classify the amino-acid pairs as predictable and unpredictable in order that we can analyse the mutation pattern in scalar data domain rather than symbol domain. Results. The results show that the mutation is highly likely to occur at the unpredictable amino-acid pairs, and the mutation has the trend to make an amino-acid pair approach predictable. Conclusion. The results provide insight on mutation from the viewpoint based on random mechanism.