The basic components of biological systems, genes, proteins, metabolites and other molecules work together in a highly orchestrated manner within cells to promote normal development and sustain health 
. Understanding how these interconnected components of biological pathways and networks are maintained in health, and how they become perturbed by genetic or environmental stressors and cause disease, is challenging but essential to developing new and better therapies to return perturbed networks to their normal state. To achieve this goal, the Library of Integrated Network-based Cellular Signatures (LINCS) project (http://lincs.hms.harvard.edu/
) aims to develop a library of molecular signatures, based on gene expression and other cellular changes that describe the response of different types of cells when exposed to various perturbing agents, including siRNAs and small bioactive molecules. Diverse high-throughput screening approaches are applied in LINCS project to interrogate the cells, which provide molecular changes and intuitive patterns (gene or protein profile) of cell response for biologists. The data acquired from these approaches were collected in a standardized, integrated, and coordinated manner 
to promote consistency and comparison across different cell types. These data will also be made openly available as a community resource that can be easily scaled up and augmented to address a broad range of basic research questions and to facilitate the identification of biological targets for new disease therapies.
Nevertheless, it is not so straightforward for biologists to uncover the cell signaling and regulatory pathways from such abundance of information. As a result, it is increasingly recognized that mathematical approaches, such as statistical inference, graph analysis, and dynamic modeling, are desired to make sense of different observed patterns. In the past decades, substantial effort has been devoted to constructing and analyzing large-scale gene or protein networks based on different types of data and literature mining. Woolf et al. 
used Bayesian approach to infer the signaling network responsible for embryonic stem cell fate responses to external cues based on measurements of 28 signaling protein phosphorylation states across 16 different factorial combinations of stimuli. The inferred network predicted novel influences between ERK phosphorylation and differentiation as well as between RAF phosphorylation and differentiated cell proliferation. The graph analysis, alone or combined with additional information regarding the network nodes, such as the functional annotation of the corresponding genes or proteins, provide testable biological predictions on several scales, from single interactions to functional modules. The functions of unannotated proteins can be inferred on the basis of the annotation of their interacting partners, as it was done for S. cerevisiae and Arabidopsis proteins using interaction, co-expression, and localization data 
. A dynamic model that correctly captures experimentally observed normal behavior allows researchers to track the changes in the system’s behavior due to perturbations. Heinrich et al. 
developed a mathematical theory that described the regulation of signaling pathways as a function of a limited number of key parameters. They found that phosphatases had a more pronounced effect than kinases on the rate and duration of signaling, whereas signal amplitude was controlled primarily by kinases. Morris et al. 
proposed a novel approach, termed constrained fuzzy logic, to convert a prior knowledge network into a computable model. Then a context-specific network model can be created by training this model against the experiment data. These models shed light on the design principles of biological control systems but are rarely context specific and cannot be used to predict the responses of cell signaling proteins as well as phenotypes to specific ligands or compounds.
In cellular pathways, especially those involved in signal transduction, kinases are known to be the major regulators which can modify up to 30% of all human proteins. Deregulation of specific kinase activity has emerged as a major mechanism by which cancer cells evade normal physiological constraints on growth and survival. As a result, there are considerable efforts to develop selective small molecule inhibitors for a host of kinases that are implicated in different cancers 
. In LINCS project, small molecule kinase inhibitors are also an area of focus in various perturbing agents tested. Thus, integration of different types of datasets in LINCS database will be a desired but challenge task to reveal the response of biological network perturbed by various kinase inhibitors. In this paper, we proposed a novel strategy to study the kinase inhibitor induced network signatures and assess the kinase inhibitor effect by considering both suppression effect on cancer cells and side effects on primary human hepatocytes in silico
. This strategy integrated KINOMEscan data of kinase inhibitors, proliferation/mitosis imaging data of cancer cells and cue signaling response data of primary human hepatocytes in LINCS database to establish a systematical network model. To our knowledge, proliferation/mitosis imaging data was first used to establish the content-specific pathway model which can bridge the gap between specific kinase inhibitors and cell phenotypes. PC9 cell line was then chosen as an example of cancer cells for specific pathway development. Integrating this PC9 cell specific model with side effects on primary human hepatocytes, we can screen out the proper kinase inhibitors with optimal concentration levels to suppress the PC9 cancer cell proliferation while avoiding severe damage to primary human hepatocytes.