A Deep Learning Approach to Causal Inference in Human Genomics using Counterfactual Reasoning DOI Creative Commons
Tshepo Kitso Gobonamang

Published: July 17, 2023

In this paper, we delve into the intricate realm of human genomics, presenting a novel design that leverages deep learning and counterfactual reasoning for causal inference. We postulate mutations occurring within DNA sequences have potential to instigate diseases by interrupting essential biological processes, hypothesis fundamentally drives research. To test this, undertaken meticulous extraction key attributes from range databases hosted National Center Biotechnology Information (NCBI). These are subsequently processed using one-hot encoding, technique effectively transforms categorical variables form could be provided machine algorithms. A sophisticated model is then utilized ascertain accuracy hypothesis. The output, depicted as graph, elucidates relationships interactions between in question, providing graphical representation proposed Our research suggests strategic modifications sequence or alterations set induce significant changes processes. This, turn, can lead structure function proteins, cornerstone cellular operations. also underline importance statements formulating hypotheses driving intelligent behavior. Despite their untestable nature inherent subjectivity, these counterfactuals serve powerful tools comprehending predicting outcomes. implications extend beyond academic interest. It provides pathway deeper understanding genomics holds promise development targeted therapies genetic diseases. fosters possibility personalized medicine therapeutic strategies alter course disease at level, potentially revolutionizing healthcare.

Language: Английский

Comprehensive analysis of the metabolomics and transcriptomics uncovers the dysregulated network and potential biomarkers of Triple Negative Breast Cancer DOI Creative Commons

Sisi Gong,

Rongfu Huang,

Meie Wang

et al.

Journal of Translational Medicine, Journal Year: 2024, Volume and Issue: 22(1)

Published: Nov. 11, 2024

Triple-negative breast cancer (TNBC) is known for its aggressive nature, lack of effective diagnostic tools and treatments, generally poor prognosis. The objective this study was to investigate metabolic changes in TNBC using metabolomics approaches explore the underlying mechanisms through integrated analysis with transcriptomics. In study, serum untargeted profiles were first examined between 18 patients 21 healthy control (HC) subjects liquid chromatography-mass spectrometry (LC-MS), identifying a total 22 significantly differential metabolites (DMs). Subsequently, receiver operating characteristic revealed that 7-methylguanine could serve as potential biomarker both discovery validation sets. Additionally, transcriptomic datasets retrieved from GEO database identify differentially expressed genes (DEGs) normal tissues. An integrative DMs DEGs conducted, uncovering molecular TNBC. Notably, three pathways—tyrosine metabolism, phenylalanine glycolysis/gluconeogenesis—were enriched, providing insight into energy metabolism disorders Within these pathways, two (4-hydroxyphenylacetaldehyde oxaloacetic acid) six (MAOA, ADH1B, ADH1C, AOC3, TAT, PCK1) identified key components. summary, highlights biomarkers potentially be used diagnosis screening comprehensive transcriptomics data offers validated in-depth understanding metabolism.

Language: Английский

Citations

0
CancerHubs: a systematic data mining and elaboration approach for identifying novel cancer-related protein interaction hubs DOI Creative Commons
Ivan Ferrari, Federica De Grossi,

Giancarlo Lai

et al.

Briefings in Bioinformatics, Journal Year: 2024, Volume and Issue: 26(1)

Published: Nov. 22, 2024

Conventional approaches to predict protein involvement in cancer often rely on defining either aberrant mutations at the single-gene level or correlating/anti-correlating transcript levels with patient survival. These are typically conducted independently and focus one a time, overlooking nucleotide substitutions outside of coding regions mutational co-occurrences genes within same interaction network. Here, we present CancerHubs, method that integrates unbiased data, clinical outcome predictions interactomics define novel cancer-related hubs. Through this approach, identified TGOLN2 as putative broad tumour suppressor EFTUD2 multiple myeloma oncogene.

Language: Английский

Citations

0
MFI2 upregulation promotes malignant progression through EGF/FAK signaling in oral cavity squamous cell carcinoma DOI Creative Commons

Wei‐Chen Yen,

Kai‐Ping Chang, Cheng‐Yi Chen

et al.

Cancer Cell International, Journal Year: 2023, Volume and Issue: 23(1)

Published: June 12, 2023

Abstract Oral squamous cell carcinoma (OSCC) is the predominant histological type of head and neck (HNSCC). By comparing differentially expressed genes (DEGs) in OSCC-TCGA patients with copy number variations (CNVs) that we identify OSCC-OncoScan dataset, herein identified 37 dysregulated candidate genes. Among these potential genes, 26 have been previously reported as proteins or HNSCC. 11 novel candidates, overall survival analysis revealed melanotransferrin (MFI2) most significant prognostic molecular patients. Another independent Taiwanese cohort confirmed higher MFI2 transcript levels were significantly associated poor prognosis. Mechanistically, found knockdown reduced viability, migration invasion via modulating EGF/FAK signaling OSCC cells. Collectively, our results support a mechanistic understanding role for promoting invasiveness OSCC.

Language: Английский

Citations

1
A Deep Learning Approach to Causal Inference in Human Genomics using Counterfactual Reasoning DOI Creative Commons
Tshepo Kitso Gobonamang, Dimane Mpoeleng

Published: July 17, 2023

<p>In this paper, we delve into the intricate realm of human genomics, presenting a novel design that leverages deep learning and counterfactual reasoning for causal inference. We postulate mutations occurring within DNA sequences have potential to instigate diseases by interrupting essential biological processes, hypothesis fundamentally drives research.</p> <p>To test this, undertaken meticulous extraction key attributes from range databases hosted National Center Biotechnology Information (NCBI). These are subsequently processed using one-hot encoding, technique effectively transforms categorical variables form could be provided machine algorithms.</p> <p>A sophisticated model is then utilized ascertain accuracy hypothesis. The output, depicted as graph, elucidates relationships interactions between in question, providing graphical representation proposed hypothesis.</p> <p>Our research suggests strategic modifications sequence or alterations set induce significant changes processes. This, turn, can lead structure function proteins, cornerstone cellular operations.</p> <p>We also underline importance statements formulating hypotheses driving intelligent behavior. Despite their untestable nature inherent subjectivity, these counterfactuals serve powerful tools comprehending predicting outcomes.</p> <p>The implications extend beyond academic interest. It provides pathway deeper understanding genomics holds promise development targeted therapies genetic diseases. fosters possibility personalized medicine therapeutic strategies alter course disease at level, potentially revolutionizing healthcare.</p>

Language: Английский

Citations

0
A Deep Learning Approach to Causal Inference in Human Genomics using Counterfactual Reasoning DOI Creative Commons
Tshepo Kitso Gobonamang

Published: July 17, 2023

In this paper, we delve into the intricate realm of human genomics, presenting a novel design that leverages deep learning and counterfactual reasoning for causal inference. We postulate mutations occurring within DNA sequences have potential to instigate diseases by interrupting essential biological processes, hypothesis fundamentally drives research. To test this, undertaken meticulous extraction key attributes from range databases hosted National Center Biotechnology Information (NCBI). These are subsequently processed using one-hot encoding, technique effectively transforms categorical variables form could be provided machine algorithms. A sophisticated model is then utilized ascertain accuracy hypothesis. The output, depicted as graph, elucidates relationships interactions between in question, providing graphical representation proposed Our research suggests strategic modifications sequence or alterations set induce significant changes processes. This, turn, can lead structure function proteins, cornerstone cellular operations. also underline importance statements formulating hypotheses driving intelligent behavior. Despite their untestable nature inherent subjectivity, these counterfactuals serve powerful tools comprehending predicting outcomes. implications extend beyond academic interest. It provides pathway deeper understanding genomics holds promise development targeted therapies genetic diseases. fosters possibility personalized medicine therapeutic strategies alter course disease at level, potentially revolutionizing healthcare.

Language: Английский

Citations

0