Research On Gene Regulatory Network Modeling Based On Fuzzy Petri Nets

Advances in techniques for high throughput data gathering such as microarrays and DNA sequencing machines have led to the integration of large-scale genome data.Transforming the enormous quantities of genomic data to useful biological knowledge is the present biggest challenge,which has opened up new research avenues in genomics.In addition,it is more important to explore the causal relationship of gene regulatory and to evaluate the function of DNA sequence.Compared with foreign countries,the level of gene regulation network in our country is still relatively backward,so as to improve the present situation,this paper presents application of Fuzzy Petri Net(FPN)in determining confidence values for bases called in DNA sequence and predicting the expression level of target gene based on the input expression level of activator/repressor gene.According to the problem of DNA sequence confidence value,a Fuzzy Colored Petri Net(FCPN)approach which combines FPN with Colored Petri Net(CPN)is proposed to model fuzzy rule-based reasoning and determine confidence values(cv)for bases called in DNA sequence.The three input features in our fuzzy model-the height,the peakness,and the spacing can be formulated as uncertain fuzzy tokens to determine the cv.The FCPN components correspond to different kinds of fuzzy operations of If-parts and Then-parts in fuzzy rules.As the FCPN model is combined by the CPN and FPN,the FCPN method not only can guarantee the quantity of information but can simplify the network size and shorten the calculation time.Thus it can provide a complete structural representation.Compared with the method of FPN,FCPN model performs more accurately which demonstrates that the proposed method is feasible and available.Aiming at the problems of gene expression level,the approach based on the Fuzzy Petri net(FPN)and reverse reasoning theory is proposed to model gene regulatory network which predicts change in expression level of the target based on the input expression level of activator/repressor.The proposed model can not only predict the expression level of target based and give causal relationship perfectly but also determine the specific win rules as well as the causal relationship easily for people.What's more,it can also determine the magnitude of activator/repressor on target gene expression level accurately.Therefore,it is helpful to research and exploration of pharmaceutical,disease diagnosis and other fields.