Identification of tissue-specific tumor biomarker using different optimization algorithms

Abstract

Background Identification of differentially expressed genes, i.e., genes whose transcript abundance level differs across different biological or physiological conditions, was indeed a challenging task. However, the inception of transcriptome sequencing (RNA-seq) technology revolutionized the simultaneous measurement of the transcript abundance levels for thousands of genes. Objective In this paper, such next-generation sequencing (NGS) data is used to identify biomarker signatures for several of the most common cancer types (bladder, colon, kidney, brain, liver, lung, prostate, skin, and thyroid) Methods Here, the problem is mapped into the comparison of optimization algorithms for selecting a set of genes that lead to the highest classification accuracy of a two-class classification task between healthy and tumor samples. As the optimization algorithms Artificial Bee Colony (ABC), Ant Colony Optimization, Differential Evolution, and Particle Swarm Optimization are chosen for this experiment. A standard statistical method called DESeq2 is used to select differentially expressed genes before being feed to the optimization algorithms. Classification of healthy and tumor samples is done by support vector machine Results Cancer-specific validation yields remarkably good results in terms of accuracy. Highest classification accuracy is achieved by the ABC algorithm for Brain lower grade glioma data is 99.10%. This validation is well supported by a statistical test, gene ontology enrichment analysis, and KEGG pathway enrichment analysis for each cancer biomarker signature Conclusion The current study identified robust genes as biomarker signatures and these identified biomarkers might be helpful to accurately identify tumors of unknown origin

A Neural Network Framework for Predicting the Tissue-of-Origin of 15 Common Cancer Types Based on RNA-Seq Data

“…Several hallmarked studies indicated that the cellular origin signatures that are expressed in primary tissue are sufficiently retained even after primary cancer cells undergo dedifferentiation and colonization in different tissue types (Ma et al, 2005;Tothill et al, 2005). A recent study compared four different algorithms and indicated that the modeling performance differed between these algorithms when analyzing RNA-Seq data from 4,127 primary tumor tissue samples related to nine cancer types (Bhowmick et al, 2019). Among those, ABC yielded the best results; it had an average precision of 91.16% and an average sensitivity of 96.5% for nine cancer types (Bhowmick et al, 2019).…”

“…Support vector machines (SVMs) based on the recursive feature elimination (RFE) algorithm represent embedded methods used for feature selection and classification modeling based on microarray gene expression data, which mined 11,925 genes to 154 genes with definite biological significance (Xu et al, 2016). More than 20,000 genes were generated from NGS RNA-Seq data in other studies (Bhowmick et al, 2019); this number is almost twice as much as that from microarray gene expression data. Hence, RNA-Seq data from nine cancer types (lung, liver, colon, thyroid, prostate, bladder, kidney, brain, and skin) were analyzed with different algorithms, and Artificial Bee Colony (ABC) yielded better results than Ant Colony Optimization, Differential Evolution, and Particle Swarm Optimization.…”

“…Hence, RNA-Seq data from nine cancer types (lung, liver, colon, thyroid, prostate, bladder, kidney, brain, and skin) were analyzed with different algorithms, and Artificial Bee Colony (ABC) yielded better results than Ant Colony Optimization, Differential Evolution, and Particle Swarm Optimization. Among different cancer types, lower grade brain glioma had the highest accuracy (99.1%) based on the ABC algorithm (Bhowmick et al, 2019). However, the robustness of feature selection and classification modeling methods still needs to be comprehensively evaluated; different algorithms might result in different results depending on their model (Chopra et al, 2010;Bhowmick et al, 2019).…”

“…Among different cancer types, lower grade brain glioma had the highest accuracy (99.1%) based on the ABC algorithm (Bhowmick et al, 2019). However, the robustness of feature selection and classification modeling methods still needs to be comprehensively evaluated; different algorithms might result in different results depending on their model (Chopra et al, 2010;Bhowmick et al, 2019). Therefore, it is necessary to design a robust classification algorithm based on NGS data that can yield accurate cancer type classification and supplement clinical examination.…”

A Prediction-based Curriculum Analysis using the Modified Artificial Bee Colony Algorithm

“…Also, the said technique fosters stability of the model, as stated in [40] that stability in variable selection technique is similarly significant to the accuracy of the prediction model. The study of [41] also a utilized majority voting scheme.…”