Reviewing methods of deep learning for intelligent healthcare systems in genomics and biomedicine

Zafar, Imran; Anwar, Shakila; kanwal, Faheem; Yousaf, Waqas; Un Nisa, Fakhar; Kausar, Tanzeela; ul Ain, Qurat; Unar, Ahsanullah; Kamal, Mohammad Amjad; Rashid, Summya; Khan, Khalid Ali; Sharma, Rohit

doi:10.1016/j.bspc.2023.105263

Cited by 21 publications

(9 citation statements)

References 219 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Deep-learning models combined with explainable artificial intelligence have potential for broad applications in precision medicine, from enhancing disease diagnosis to facilitating drug discovery [ 69 , 70 , 71 ]. Deep-learning models offer more exact and efficient diagnosis for diseases requiring analysis of medical images (i.e., cancer, dementia), compared with human experts [ 72 ].…”

Section: Resultsmentioning

confidence: 99%

“…Explainable artificial intelligence makes AI algorithms more transparent and controllable, building trust among medical professionals in AI-assisted decisions [ 78 ]. Overall, explainable artificial intelligence integration into the healthcare systems can build trust and reliance in deep-learning approaches to diagnosis and drug discovery [ 56 , 69 , 79 ].…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

The Promise of Explainable AI in Digital Health for Precision Medicine: A Systematic Review

Allen

2024

JPM

View full text Add to dashboard Cite

This review synthesizes the literature on explaining machine-learning models for digital health data in precision medicine. As healthcare increasingly tailors treatments to individual characteristics, the integration of artificial intelligence with digital health data becomes crucial. Leveraging a topic-modeling approach, this paper distills the key themes of 27 journal articles. We included peer-reviewed journal articles written in English, with no time constraints on the search. A Google Scholar search, conducted up to 19 September 2023, yielded 27 journal articles. Through a topic-modeling approach, the identified topics encompassed optimizing patient healthcare through data-driven medicine, predictive modeling with data and algorithms, predicting diseases with deep learning of biomedical data, and machine learning in medicine. This review delves into specific applications of explainable artificial intelligence, emphasizing its role in fostering transparency, accountability, and trust within the healthcare domain. Our review highlights the necessity for further development and validation of explanation methods to advance precision healthcare delivery.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

The Promise of Explainable AI in Digital Health for Precision Medicine: A Systematic Review

Allen

2024

JPM

View full text Add to dashboard Cite

show abstract

“…GSEA analysis , ADAMTS6 is full of cancer-related pathways, like VEGF, which regulates angiogenesis. Vascular endothelial growth factor signaling pathway inhibition has been demonstrated to impede cardiovascular formation, preventing the development and propagation of tumors as earlier researchers investigated [ 12 , 56 ]. This indicates that ADAMTS6 is closely related to endothelial cells [ 57 ].…”

Section: Resultsmentioning

confidence: 99%

“…With the advancement of single-cell RNA sequencing (scRNA-seq) technology in recent years, a vast amount of scRNA-seq data has been generated [ 10 ]. Researchers [ 11 ], delve into the wealth of biological insights inherent in scRNA-seq data by scrutinizing cell information and uncovering heterogeneity among cells, thereby offering valuable insights into the relationships between cells, genes, and diseases [ 12 ].…”

Section: Introductionmentioning

confidence: 99%

Cauchy hyper-graph Laplacian nonnegative matrix factorization for single-cell RNA-sequencing data analysis

Wang,

Shen

2024

BMC Bioinformatics

View full text Add to dashboard Cite

Many important biological facts have been found as single-cell RNA sequencing (scRNA-seq) technology has advanced. With the use of this technology, it is now possible to investigate the connections among individual cells, genes, and illnesses. For the analysis of single-cell data, clustering is frequently used. Nevertheless, biological data usually contain a large amount of noise data, and traditional clustering methods are sensitive to noise. However, acquiring higher-order spatial information from the data alone is insufficient. As a result, getting trustworthy clustering findings is challenging. We propose the Cauchy hyper-graph Laplacian non-negative matrix factorization (CHLNMF) as a unique approach to address these issues. In CHLNMF, we replace the measurement based on Euclidean distance in the conventional non-negative matrix factorization (NMF), which can lessen the influence of noise, with the Cauchy loss function (CLF). The model also incorporates the hyper-graph constraint, which takes into account the high-order link among the samples. The CHLNMF model's best solution is then discovered using a half-quadratic optimization approach. Finally, using seven scRNA-seq datasets, we contrast the CHLNMF technique with the other nine top methods. The validity of our technique was established by analysis of the experimental outcomes.

show abstract

“…Among the three most common DL algorithms we have (1) Multi-layer perceptron (MLP), which is particularly effective for tasks that involve non-sequential and non-spatial data; (2) Convolutional neural network (CNN) designed for visual data, such as images and videos; and (3) Transformer [7], the backbone of most large language models (LLM), used in natural language processing (NLP) tasks, suited for any data that can be represented as a sequence, including text, time-series data, and even genomic sequences. DL has been successfully applied in many domains, such as computer vision, speech recognition, text and image generation, and biomedicine [8, 9, 10, 11, 12].…”

Section: Introductionmentioning

confidence: 99%

Predicting nonlinear genetic relationships between traits in multi-trait evaluations by using a GBLUP-assisted Deep Learning model

Shokor,

Croiseau,

Gangloff

et al. 2024

Preprint

View full text Add to dashboard Cite

Background: Genomic prediction aims to predict the breeding values of multiple complex traits assumed to be normally distributed, thus imposing linear genetic correlations between traits. However, these statistical methods are unable to model nonlinear genetic relationships between traits, if existent, potentially leading to a decrease in prediction accuracy. Deep learning (DL) is a promising methodology for predicting multiple complex traits, in scenarios where nonlinear genetic relationships are present, due to its capacity to capture complex and nonlinear patterns in large data. We proposed a novel pure DL model, designed to obtain predicted genetic values (PGV) while accounting for nonlinear genetic relationships between traits, and extended this model to a hybrid DLGBLUP model which uses the output of the traditional GBLUP, and enhances its PGV by using DL. Using simulated data, we compared the accuracy of the PGV obtained with the proposed pure DL model, the hybrid DLGBLUP model, and the traditional GBLUP model, the latter being our baseline reference. Results: We found that both DL and DLGBLUP models either outperformed GBLUP, or presented equally accurate PGV, with a particular greater accuracy for traits presenting a strongly characterized nonlinear genetic relationship. DLGBLUP presented the highest prediction accuracy and smallest mean squared error of the PGV for all traits. Additionally, we evolved a base population over seven generations and compared the genetic progress when selecting individuals based on the additive PGV obtained by either DL, DLGBLUP or GBLUP. For all traits with a nonlinear genetic relationship, after the fourth generation, the observed genetic gain when selection was based on the additive PGV from GBLUP was always inferior to the observed when selection was based on either DL or DLGBLUP. Conclusions: The integration of DL into genomic prediction has potential to bring significant advancements in the field. By identifying nonlinear genetic relationships, our DL and DLGBLUP models improved prediction accuracy. It offers an insight to genetic relationship and its evolution over generations, with potential to improve selection strategies in commercial livestock breeding programs. Moreover, DLGBLUP shows that DL can be used as a complement to statistical methods, by enhancing their performance.

show abstract

Reviewing methods of deep learning for intelligent healthcare systems in genomics and biomedicine

Cited by 21 publications

References 219 publications

The Promise of Explainable AI in Digital Health for Precision Medicine: A Systematic Review

The Promise of Explainable AI in Digital Health for Precision Medicine: A Systematic Review

Cauchy hyper-graph Laplacian nonnegative matrix factorization for single-cell RNA-sequencing data analysis

Predicting nonlinear genetic relationships between traits in multi-trait evaluations by using a GBLUP-assisted Deep Learning model

Contact Info

Product

Resources

About