Metabolomics the New Diagnosis of Breast Cancer
Prasanna Priya Golagani1, Shaik Khasim Beebi2, Tummala Sita Mahalakshmi2
1Department of Computer Science and Engineering, VIGNAN IIT (Autonomous), VSP, AP, INDIA.
email id: [email protected]
2Department of Biotechnology, GITAM University, VSP, AP, INDIA.
3Department of Computer Science and Engineering, GITAM University,VSP,AP,INDIA.
ABSTRACT
Chest Carcinoma is one of the leading causes of death in women. Early prediction of chest carcinoma can bring up the chances of viability. Chest carcinoma can be diagnosed by a numerous variety of tests which include a mammogram, ultrasound, MRI and biopsy. 90% of them are image based. Cancer can also be predicted much before a lump is formed by studying the metabolism of the breast cancer cell. The major divergence between a normal breast cell and a breast cancerous cell is that the apoptotic cycle and the transcriptional cycle are malfunctioning. This malfunction is caused by over or under production of certain metabolites. In our work we identified 8 such metabolites and their correlated metabolic cycles. Reactome and KEGG databases are used to study these pathways in detail. Computational models are developed for these 8 pathways in Cell Designer and the metabolomics of these pathways are collected from different curated data bases such as BioModels.net and they are simulated or executed using Control Panel and Copasi GUI. The metabolomics of these computationally modelled pathways are developed into a dataset.
Keywords
Metabolomics; Chest Carcinoma; Breast Cancer Cell; Kegg; Cell Designer
Introduction
Omics is a branch of knowledge of life sciences. Omics integrates genomics, proteomics, transcriptomics and metabolomics. Metabolomics integrates the methods to locate and calculate cell metabolites with the help of complicated and logical techniques by using numerical and numerous varieties of procedures for data unsheathing and information elucidation. The metabolomics provides the complete information regarding the physical state of a human being. Mass spectrometry along with liquid, gas or NMR chromatography can be used to audit numerous metabolites concurrently. [1]
1.1. Reactome:
REACTOME is a free open source curated and peer reviewed human pathway knowledge base.
It contains normal and disease-related pathways. Every molecule and every event in REACTOME are mapped with the correct cellular compartments. . This database is called REACTOME because reactions are the steps involved in the pathways, they can either be binding, dissociation, degradation, phosphorylation, dephosphorylation, transport etc.
1.2. KEGG
KEGG stands for Kyoto Encyclopedia of genes and genomes. KEGG is a combined database of 18 handpicked databases. These 18 databases are categorized into 4 groups. KEGG Pathway database is a group of physically worn track plots showcasing understanding of the elementary reciprocity, retort and association webwork for Metabolism, Genetic Information, Human Diseases, and Drug Development etc. Human diseases database consists pathways related to Cancer overview, Cancer specific types, Immune diseases, etc. Cancer of specific types contains pathways related to Thyroid cancer, Basal cell carcinoma, Breast cancer etc. In Breast cancer there are pathways related to Luminal A breast cancer, Luminal B breast cancer, HER2 positive breast cancer, Basal like/Triple negative breast cancer.
1.3. Cell Designer
Cell Designer is a structured diagram editor for drawing gene-regulatory and biochemical networks that uses standard formats. It provides a user-friendly interface for graphical, biochemical pathway description. Cell Designer is used to visualize the modelling networks.
Apart from SBML Cell Designer also supports other standard formats like BioPax and SBGN.
Cell Designer sustains graphical notations, mathematical simulations and database connections. The models developed in Cell Designer are simulated using Control Panel and
1.4. SBML
Systems Biology Markup Language is a programming language that can be used for simulation of a breast cancer pathway. SBML allows models to be encoded using xml. SBML can be used as a format for reading and writing models. It can also be used by different software tools for building and editing models, simulation of programs and the databases can directly communicate and store the same. This allows sharing results and permits other researchers to start with an unambiguous representation of the model; they can examine it carefully, propose precise corrections and extensions and apply new techniques and approaches to do better science.
1.5. Biomodels
Biomodels is a depot of arithmetic designs showcasing organic structures. Presently is converges designs under metabolic pathways, signalling, amino acid- medicine interplay, pandemic models and several more. It permits, depositing, finding and recollecting the arithmetic designs. Biomodels has quantitative data showcasing the metabolites concentration and the activity of the design.
Literature Review
Hosseini et al (2020) and Hassanpour got stimulated by the fact that there is no importance given to the metabolomics and their role in metabolic pathways. Inspite of the progress in anticipating a group of quantification their part in the life sciences is still a question to take. To anticipate the quantifications, they developed a hypothesizing supported method. They termed it PUMA (Probabilistic modeling for Untargeted Metabolomics Analysis). It identifies metabolomics quantification and the organic pathways from the organic specimen which is studied in a creative design with a hypothetical specimen to calculate anticipated allocation.
Since there is no Metabolomics data sets where all quantities are elucidated and the track functioning is familiar. They certified PUMA on artificial datasets that are modelled to imitate cellular processes.[2]
Marcelo et al (2019) and Hector tried to validate between the chromosome data and the pathway data so they built designs that are grounded to chromosomes separately and grounded to specimen related results for every pathway. They reduced the quality of the information by putting disturbance into it and scrutinized the capacity of the designs to give correct output.
They found that the designs in the pathway space has greater accuracy than the designs in the chromosome space. [3]
Antonio et al (2018) and Florian identified that the pathway databases use conventional directory for saving their data. This cuts down on the retrieval capacity because there is difficulty in passing over the questions in greatly interdependent information. They found out that grid database works better, it can be questioned better or data can be retrieved easily, for
the same information. So, a graph database can be used and for retrieving data subject assistance can be used over conventional directory system. [4]
Paul et al (2017) and Fintan identified that studying the biological pathways is important to cure illnesses like carcinoma. Optical presentation of these pathways helps researchers to study them easily and better. So, they put forward a classification for the functions to be done to obtain organic pathway information. They collected these functions by discussing the process of studying a pathway with several professionals. They also studied the optical techniques that help these functions. [5]
Yvan et al (2016) and Sofie conducted a review on the current progress in flow cytometry which helps researchers to calculate the escalating figure of specifications for cell, creating a great number of datasets. To scrutinize, envision and clarify this information, recently accessible software approaches must be used to solve and improvise, by the researchers.
Computational flow cytometry is evolving as a field of scope and interconnection of physiology and computational life sciences. [6]
Irene et al (2015) and Anika put together the importance of capturing quick connection between chromosome manifestation, pathways and physical traits, which is very important to examine and study illnesses and create medicines. There are numerous tasks taking place to connect the chromosomes, chromosome utterance, design pathways and categorize physical traits. A number of databases were developed and numerous softwares for putting together these data sets and making a interlink between them were tried. In this paper they discussed the areas which can be examined to for better amalgamation of chromosomes, pathways and phenotypes and this helps in better understanding how chromosome aberrations change pathways and how they in turn change the physical traits.[7]
Wim et al (2014) and Henk proposed a new method to identify communicating pathway action grounded on education grounded Bayesian software designs, which translate measured protein synthesis information as a methodical result of a dynamic communicating pathway, by using manifestation measure of protein synthesis chromosomes. After standardization on less cells, they can be used to validate carcinomas of tissues or organs, they presented Wnt and ER pathways with the laboratory assessment on individual datasets from various cancer types.
They found out that Wnt pathway was involved in thirty percent of Chest carcinoma and ER pathway was involved in thirty nine percent of chest carcinoma.[8] Liana et al (2013) and David scrutinized the chromosome utterance outline in chest carcinoma sufferers. They selected a data set which contains the protein synthesis outline of 570 homo sapiens chest carcinoma.
Using Machine Learning they tried to locate metabolite utterance related to illness handling pathways and biological marker therapy blueprint related to beta 2 adrenergic receptor cell signaling. They identified a group of chromosomes which are the cause for carcinoma and also have a role in beta adrenergic receptor signaling.[9] Ke-Qin Liu et al (2012) and Zhi-Ping designed a latest way to build pathway interconnection webwork grounded on chromosome utterance, amino acid amino acid interconnections and metabolic pathways. They noticed the variations in the pathways in carcinoma and targeted them. The variations in the pathway are treated as a sub category under the category of interconnection of the pathways, these sub categories expose very well the interaction between the pathways. They run this design on several carcinoma datasets. The standard outputs of several carcinoma datasets illustrate that this design is having a scope for future advancements in this area.[10]
Robert and Ayesha explained the major provocations paving webwork designing of endocrine resistance, a large number of them come up from the characteristics of the information gaps that are being scrutinized. When provocation of evolved amino acid retaliation is used along
with ordination of ergastoplasm strain in chest carcinoma units by antiestrogens to demonstrate how computational modeling is done. Provocation of evolved amino acid retaliation is the main factor for the future of the cell to deicide and start degrading of cell and cell death. [11]
Proposed Work
Breast cancer has been diagnosed by a mammogram, ultrasound, MRI and biopsy until now.
Ninety percent of them are image based. Cancer can be predicted much before a lump is formed by studying the metabolism of the breast cancer cell. The main difference between a normal breast cell and a breast cancerous cell is that the apoptotic cycle and the transcriptional cycle are malfunctioning. This malfunction is caused by over or under production of certain metabolites. In our work we identified 8 such metabolites and their correlated metabolic cycles. Computational models are developed for these 8 pathways which are Protein Kinase B, Mitogen Activated Protein Kinase B, Mammalian Target Of Rapamycin, Type II Trans membrane protein, Single Pass Transmembrane Receptor, Sonic Hedgehog, Tumor Necrosis Factor, Wingless/Integrated Pathways in Cell Designer [12]. Metabolomics of these pathways are collected from different curated data bases such as BioModels.net and they are simulated using Control Panel and Copasi GUI in the Cell Designer. The metabolomics of these computationally modelled pathways are developed into a dataset. The Data set is a labelled dataset, which contains the concentrations of the 8 signal transduction pathways. High concentrations of the metabolites cause breast cancer. The label specifies whether the metabolite values of the tuple cause cancer or not.
Results
Path c1 c2 c3 c4 c5 c6 c7 c8 c9
yes /no AKT 5.5 1.5 1.5 1.5 2.5 1.5 3.5 1.5 1.5 no AKT 5.5 4.5 4.5 5.5 7.5 10.5 3.5 2.5 1.5 yes FASL 10.5 6.5 4.5 1.5 3.5 4.5 3.5 2.5 3.5 yes FASL 1.5 1.5 2.5 1.5 2.5 2.5 4.5 2.5 1.5 no MAPK 8.5 8.5 7.5 4.5 10.5 10.5 7.5 8.5 7.5 yes MAPK 1.5 1.5 1.5 1.5 1.5 1.5 3.5 1.5 1.5 no NOTCH 10.5 5.5 5.5 6.5 3.5 10.5 7.5 9.5 2.5 yes NOTCH 1.5 1.5 1.5 1.5 2.5 1.5 2.5 1.5 1.5 no SHH 9.5 1.5 2.5 6.5 4.5 10.5 7.5 7.5 2.5 yes SHH 8.5 4.5 10.5 5.5 4.5 4.5 7.5 10.5 1.5 yes TNF 3.5 1.5 2.5 1.5 2.5 1.5 2.5 1.5 1.5 no TNF 1.5 1.5 1.5 1.5 2.5 1.5 1.5 1.5 1.5 no
WNT 5.5 7.5 9.5 8.5 6.5 10.5 8.5 10.5 1.5 yes
WNT 4.5 1.5 1.5 3.5 1.5 1.5 2.5 1.5 1.5 no MTOR 3.5 1.5 1.5 1.5 2.5 1.5 2.5 1.5 1.5 no MTOR 3.5 1.5 1.5 2.5 2.5 1.5 1.5 1.5 1.5 no
Table 1. Metabolomics Data Set of Breast Cancer
Figure 1: The Simulated WNT (Wingless/Integrated) pathway using Control Panel in the Cell Designer.
Figure 2: The Simulated WNT pathway using Control Panel in the Cell Designer.
Figure 3: The Simulated WNT pathway using COPASI GUI in the cell designer.
Figure 4: The Simulated NOTCH (Single Pass Transmembrane Receptor) pathway using Control Panel in the Cell Designer.
Figure 5: The Simulated NOTCH pathway using Control Panel in the Cell Designer.
Figure 6: The Simulated Notch pathway using COPASI GUI in the cell designer.
Discussion
The Metabolomics of a breast cancer cell and a healthy breast cell is taken and developed into a data set. Cell Metabolomics has great scope in future to study a disease and develop a medicine for it. This data set is a labeled data set. The label specifies weather the cell has cancer or not. This data set is developed by taking the data related to 8 signal transduction pathways into account. Breast Cancer Prediction with cell metabolomics causes early prediction of cancer even before we can see a lump formed in the images. Metabolomics is quite promising, as it gives the complete functionality of the cell. Any malfunction can be predicted very early and this helps in increasing survivability of the patient.
Conclusion
Metabolomics is the future of Breast Cancer prediction. Early breast cancer prediction is a challenge that we are facing today and metabolomics aids in early prediction of breast cancer.
Studying the cell gives us the whole idea of the tissue or organ and this helps in predicting other health issues as well apart from cancer.
Future Work
Machine Learning is a branch of Artificial Intelligence, which uses the data set to make a machine, learn like a human brain. Supervised learning algorithms can be used as this is a labeled data set. Using python programming we can develop a neural network and use TensorFlow to predict the accuracy of the dataset.
References
[1] What is metabolomics all about? By Ute Roessner and Jairus Bowne published online BioTechniques, VOL. 46, NO. 5 doi.org/10.2144/000113133
[2] Pathway-Activity Likelihood Analysis and Metabolite Annotation for untargeted Metabolomics Using Probabilistic Modeling by Ramtin Hosseini, Neda Hassanpour Metabolites 2020 published May 2020
[3] Predictive modelling using pathway scores: robustness and significance of pathway collections by Marcelo, Hector Springer BMC Bioinformatics 20, Article Number:
543 (2019) is cited by 2.
[4] Reactome graph database: Efficient access to complex pathway data by Antonio, Florian PLOS Computational Biology published Jan 2018 journal. pcbi.1005968 Is cited by 145
[5] A taxonomy of visualization tasks for the analysis of biological pathway data by Paul, Fintan BMC Bioinformatics 18 Article no. 21 Feb 2017 is cited by 23.
[6] Computational flow cytometry: helping to make sense of high- dimensional immunology data by Yvan and Sofie Nature Review Immunology 16, 449-462 (2016) is cited by 186.
[7] Linking gene expression to phenotypes via pathway information by Irene, Anika Springer Journal of Biomedical Semantics 6, Article number: 17 (2015) is cited by 23.
[8] Selection of Personalized Patient Therapy through the use of Knowledge-Based Computational Models that Identify Tumor Driving Signal Transduction Pathways by Wim, Henk Cancer Research CAN 13 2515 June 2014 is cited by 23.
[9] Biomarker identification in breast cancer beta adrenergic receptor signaling ang pathway to therapeutic response by Liana, David Elsevier Computational and Structural Biotechnology Journal Volume 6 Issue 7 March 2013. Is cited by 13.
[10] Identifying dysregulated pathways in cancers from pathway interaction networks by Ke-Qin Liu, Zhi-Ping Liu BMC Bioinformatics 2012, 13:126 is cited by 115.
[11] Endoplasmic reticulum stress, the unfolded protein response, and gene network modeling in antiestrogen resistant breast cancer by Robert Clarke, Ayesha De Gruyter Hormone Molecular Biology and Clinical Investigation Volume 5 March 2011. Is cited by 46.
[12]Computational modeling of signal transduction pathways in breast cancerous cell and target therapy Golagani Prasanna Priya, Shaik Khasim Beebi, T. Sita Mahalakshmi International Journal of Recent Technology and Engineering, 2019, 8(1), pp. 1170–
1175
