China
Africa
Explore chapters and articles related to this topic
Explainable Artificial Intelligence: Guardian for Cancer Care
Published in Mohamed Lahby, Utku Kose, Akash Kumar Bhoi, Explainable Artificial Intelligence for Smart Cities, 2021
On the genetic/molecular level, the progressive accumulation of mutations is the major source of cancer initiation. But still, all the mutations do not equally contribute to cancer, as only a few of them are rare and disease drivers. The human genome contains around 518 protein kinase genes that are collectively known as kinome. Multiple classifier systems were used to segregate the mutations in positive and negative classes. Positive and negative mutations were retrieved from the ‘Catalogue of Somatic Mutations in Cancer’ (COSMIC) database and SNP@Domain database. Also, protein characteristics concerning point mutations were determined from various sources, including KinBas, Uniprot, EMBOSS, etc. Eleven machine learning methods, including J48 (Tree), Random Forest, NB Tree, Functional Tree, Decision Table, DTNB, LWL (J48+KNN), Bayes Net, Naive Bayes, SVM, and Neural Network, were applied for the development of classifier system to identify EGFR mutations. For the first time, the connection between EGFR mutations T725M and L861R was established with cancers (U et al., 2014).
Biochemical pathways
Published in Christian Mazza, Michel Benaïm, Stochastic Dynamics for Systems Biology, 2016
Christian Mazza, Michel Benaïm
Cells receive information from their environment, and, in turn, respond in a way that is coded by their genes and epigenetic factors. Cells are able to respond to many chemical and physical agents which can induce transitory or permanent changes in cells. These signalling molecules bind to protein receptors, which interact with special proteins in the cytoplasmic or plasma membrane. The latter then transduce (or send) the signal to deeper levels within the cell. Protein kinases and protein phosphatases mediate a significant part of the signal transduction in eukaryotic cells. The human kinome contains more than 500 types of protein kinase (see, e.g., [118]) and approximatively 150 types of protein phosphatase. These pathways are designed to elicit a cellular response like the activation of transcription factors in response to external signals, and thus increase the expression of some target genes, producing in this way mRNA flux; see, e.g., [120] for a nice exposition of cellular processing, from the point of view of systems biology. The network structures associated with these protein or signalling networks are similar to the topologies observed in metabolic networks, and contain positive and negative feedback loops; see, e.g., [135]. Cellular processing units are designed and work like electronic circuits or neural networks; see, e.g., chapter 6 of [4], where such pathways are studied using tools from neural network theory. A basic difference is that the living processing units involved can move and diffuse inside the cell. It is difficult to give a general overview of the various existing topologies. The reader can consult, among others, [52], [191], [13], [177], [116], [53] and [151] to get ideas on known network structures.
Systems toxicology approach explores target-pathway relationship and adverse health impacts of ubiquitous environmental pollutant bisphenol A
Published in Journal of Toxicology and Environmental Health, Part A, 2022
Manigandan Nagarajan, Gobichettipalayam Balasubramaniam Maadurshni, Jeganathan Manivannan
The kinome wide interaction profile of BPA was predicted by employing the KinomeX tool which implements comprehensive profiling on kinome-wide activity for small molecules, and might can depict overall selectivity and selectivity toward a sub-family of kinases based upon the predicted kinase profile (Li et al. 2019). In particular, kinome wide target screening is considered as more efficient approach toward the discovery of new kinase inhibitors (Miduturu et al. 2011) and characterizing systems-level drug action (Vidović et al., 2014). In the current study, approximately 17 kinases were predicted as possible targets of BPA. Among these, the eIF2α kinase heme-regulated inhibitor (HRI) exhibited the highest interaction score which is known to protect the host from infection by regulating intracellular pathogen trafficking (Bahnan et al. 2018). Subsequently, the receptor-interacting serine/threonine-protein kinase 1 (RIPK1) is a key mediator of cell death and inflammation (Mifflin, Ofengeim, and Yuan 2020) and its inhibitors may possess significant potential for treatment of inflammatory disorders and cancer metastasis (Hou et al. 2019). Further, several investigators suggested that mammalian PAS domain-containing serine/threonine-protein kinase (PASK) is involved in glucose metabolism, insulin secretion and blood glucose levels by its action on pancreatic islet α/β cells and glycogen synthase (GS), in which the PASK knockout mice (PASK-/-) are protected from obesity, liver triglyceride accumulation and insulin resistance (Zhang et al. 2015). Our study, focused on casein kinase 2 alpha 1(CSNK2A1) and casein kinase 2 alpha 2(CSNK2A2,) catalytic subunits of casein kinase II (CSNK2) which is a pleiotropic serine/threonine kinase participates in diverse cellular processes. In this regard, recent evidence suggested that CSNK2A1 plays important oncogenic roles in gastric cancer invasion via epithelial-mesenchymal transition (EMT) and the PI3K-Akt-mTOR signaling pathways (Jiang et al. 2019). In addition, a genome-wide association study reported the association of CSNK2A2 variants with leukocyte telomere length (Saxena et al. 2014). Further, the interaction of BPA with transforming growth factor-β (TGF-β receptor) I (where phosphorylation of a Gly-Ser regulatory region observed during activation) enabled us to consider transforming growth factor-β (TGF-β) signaling also as a potential target of BPA. Since it is a pleiotropic cytokine elicits complex effects in cells that plays key role in cardiovascular physiology, hemostasis, blood-vessel interface and malignancies (Kubiczkova et al. 2012; Redondo et al. 2012).