The Mechanism Research Of Multiple Myeloma Signaling Pathways Based On The Modularized Factor Graph Model

Systems biology is a subject that can quantify and reveal the functions of biological systems.It can solve a lot of biological problems by establishing the mathematical model based on combinations of the mathematical theory and computer tools.It has been considered to be an effective strategy of exploring cancer pathogenesis with its application and development in the study of disease.This paper mainly studies the regulatory mechanism of the multiple myeloma development by using systems biology.Due to its complex pathogenic factor and low cure rate,multiple myeloma is always a big problem for human beings.Along with the improvement of clinical and biological experiment technology,it has been discovered that the hematopoietic microenvironment plays an important role in the pathogenesis and development of multiple myeloma.Especially,the signaling pathways between myeloma initial cells and bone marrow stromal cells within microenvironment affect proliferation and apoptosis of MM cells.The treatment delivered at the key targets of pathways can help MM cells recover to a certain degree and help patients get certain clinical remission.However,when it comes to studying the large-scale and complex signaling pathway,Experiment methods is hard to satisfy the new requirement.Therefore,how to identify and construct signaling pathways of multiple myeloma will be of great significance for researchers to study the disease mechanism.In the meanwhile,how to build effective calculation model for complex signaling pathway is always the focus.As we know,a lot of unknown parameters exist because of the complexity of biological systems.Hence,how to choose the right number of control parameters and how to efficiently estimate the parameter values remains a problem to be solved.To address the above two problems,we proposed a modular mechanism based on factor graph algorithm to find the target of multiple myeloma signaling pathways.This method not only finds out the key protein of multiple myeloma signaling pathway,but also effectively reduces the computation cost and improves the computational efficiency.In a word,it is important to qualitatively explore the interaction between the components within the pathway and quantify the influence on each other.Systems biology model plays a crucial role in this process.The main methods are as follows:(1)Construction of MM related signaling pathway.Firstly,we got RPPA data by using protein chip.Then we roughly screened differentially expressed proteins.Based on the initial screening results,a lot of signaling pathways for MM were identified by combining with the signaling pathway database and related studies.Finally,we selected and integrated these multiple myeloma signaling pathways into a large pathway.Due to the complexity of signaling pathways,it is difficult to model for the large pathway.Hence we simplified it according to some rules and gained the specific signaling pathways.(2)Construction of computational model of signaling pathway.According to the simplified signaling pathway,we first decomposed the whole pathway into 19 smaller modules and represented various interaction cascade of signaling pathways by using the hybrid functional Petri net and ordinary differential equations(ODEs)model based on mass action equations.It reached a good balance between clear representation of complex signaling pathways and high simulation efficiency.Then we applied PSO algorithm on individual module in turns to get the initial estimates through initial scope and got the output of large scopes.Next,we decompose the whole pathway into 2 sub-pathway based on the specific rule and constructed factor graph for each sub-pathway and define fitness function for each factor node in each factor graph based on experimental data.After that,we applied max-product BP algorithm on the large factor graph to get the output of small scopes for all parameters.This process solves conflict parameters.BP algorithm is regarded as an approach to reconcile this kind of conflict and presented as messages updating equations on factor graph.Finally,we got the final estimation within the small scope by PSO algorithm.Compared with general optimization algorithm,the novel model optimization algorithm is more stable and has less error.(3)Parameter analysis and experimental verification.In this paper,parameter uncertainty analysis and parameter sensitivity analysis were adopted to estimate the stability of the model.And parameter variation analysis has been done to explore the key target protein of multiple myeloma in signaling pathway.,FAK,mTORC1,NF?B have been singled out.Then we combined immunohistochemical experiment and related research to verify the accuracy of the model.This paper picked out the signaling pathway of myeloma and constructed a systems biology model based on experimental data.The method quantifies the combinatorial effects of interaction between signaling pathways.In conclusion,by improving the control mechanism in the process of tumor growth can improve the clinical management of diseases and help doctors diagnose and treat diseases.Thus,unnecessary chemotherapy may be avoided and the pain of the patients may be reduced.Which provide a new insight into this kind of problem and lay a foundation for further theoretical research...