Researches On Alignment Algorithm For Revealing Gene Regulatory Networks Differentiation

Many major diseases,including various types of cancer,increasingly threaten human health,yet their underlying pathogenesis remains one of the major problems facing biomedicine today.Complex biological networks provide a holistic perspective of the dynamic process of life phenomena through the eyes of systems science.Researches for differentials in complex systems focus on the key modules which result in significant changes of structures and functions between the two networks,named as Kernel Differential Sub-graph(KDS).Based on the comparison of different networks and the goal of KDS mining,this paper mainly focuses on two aspects:(1)During the developing of life process,the changes of biological networks are reflected not only in the change of topological structure,but also in the disappearance of certain biomolecules and the appearance of some new biomolecules.(2)The differentials of complex biological networks must not only involve the topological changes,but also involve the biological significance of the molecules.In this paper,we propose a new method contributed to exploring the kernel differential sub-graph of networks with different nodes(KDS-DN)based on differential network comparison,and successfully applied to the alignment of Non-small Cell Lung Cancer(NSCLC).The main works and innovations of this paper are as follows:(1)A new KDS extraction algorithm KDS-DN is proposed.The algorithm realizes the comparison of networks with a certain similarity.The number of nodes in these networks can be different,but some of the nodes are the same or similar.KDS-DN matches the nodes one by one through the similarity comparison algorithm.For the same node,the topological change theory is used for calculation of the topological change;for different nodes,both the topological changes and the biological significance of the nodes are considered and thus KDS-DN select the kernel differential nodes.Furthermore,this paper presents three principles for extracting the KDS: sub-graph connectivity,minimization of size and maximization of differential.According to these three principles,the module of core differences between networks is extracted,and the validity of KDS-DN is proved by statistical method and shortest path based method.(2)KDS-DN is successfully applied in the differential comparison of NSCLC gene regulatory networks.The preliminary work of NSCLC gene regulatory network differential comparison mainly includes standardization,gene screening,and network construction in different states.Based on differential comparison s of cancer and normal networks,30 genes most likely to be associated with NSCLC are found and the KDS modules of the relevant networks are extracted.At the same time,it also excavates two important functional modules related to PI3 K / AKT,MAPK and PRKC signaling pathways.Gene Ontology(GO)enrichment analysis and literature mining show that KDS and two functional modules excavated in this paper have a strong relationship with pathogenicity of NSCLC.In addition,KDS-DN is also proved to be reasonable and effective compared with the existing equivalent methods.Moreover,KDS-DN has the potential to predict other critical disease-related genes and modules.