The Impact of Data Deviation in Inferring a Biological Dynamic System

The high-throughput measurement devices for DNA, RNA, and proteins produce large amount of temporal data from biological dynamic systems. It is a need to conduct reverse engineering to these data to reveal parameters/ structure and outcome relationships implied in the data. We argue that, since the data available in reality are not perfect, the result of the reverse engineering is impacted by the un-perfect data. If this is true, then what is the degree of the impact? An intuitive method to study this problem is to mutate an original measured temporal data to a noise-temporal data we wanted and then employ reverse engineering to it. However, this method is not working because there is no flawless reverse algorithm we known can use. In such a case, we cannot tell the variation in the result is come from data or algorithm itself. We have to solve it in another way. We develop a novel method to investigate this problem and choose the parameter estimation as our task of reverse engineering. The basic idea is to search parameters in the parameter space that can produce outcome which has a minor deviation from the original outcome. These parameters we found can be regarded as the parameters found from a flawless reverse engineering algorithm with a minor temporal data. We then can compare these parameters with its target parameters to know the impact of the data deviation. One artificial system is used as test bed to demonstrate our approach. The results of the experiments show, a minor deviation in data may introduce a large deviation in the parameter solutions. We conclude that we should not ignore the effect of data deviation in reverse engineering even it is just a minor deviation.