Targeted deep learning classification and feature extraction for clinical diagnosis
Summary: Protein biomarkers can be used to characterize symptom classes, which describe the metabolic or immunodeficient state of patients during the progression of a specific disease. Recent literature has shown that machine learning methods can complement traditional clinical methods in identifyin...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2023-11-01
|
| Series: | iScience |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2589004223020837 |
| _version_ | 1797657861249040384 |
|---|---|
| author | Yiting Tsai Vikash Nanthakumar Saeed Mohammadi Susan A. Baldwin Bhushan Gopaluni Fei Geng |
| author_facet | Yiting Tsai Vikash Nanthakumar Saeed Mohammadi Susan A. Baldwin Bhushan Gopaluni Fei Geng |
| author_sort | Yiting Tsai |
| collection | DOAJ |
| description | Summary: Protein biomarkers can be used to characterize symptom classes, which describe the metabolic or immunodeficient state of patients during the progression of a specific disease. Recent literature has shown that machine learning methods can complement traditional clinical methods in identifying biomarkers. However, many machine learning frameworks only apply narrowly to a specific archetype or subset of diseases. In this paper, we propose a feature extractor which can discover protein biomarkers for a wide variety of classification problems. The feature extractor uses a special type of deep learning model, which discovers a latent space that allows for optimal class separation and enhanced class cluster identity. The extracted biomarkers can then be used to train highly accurate supervised learning models. We apply our methods to a dataset involving COVID-19 patients and another involving scleroderma patients, to demonstrate improved class separation and reduced false discovery rates compared to results obtained using traditional models. |
| first_indexed | 2024-03-11T17:50:42Z |
| format | Article |
| id | doaj.art-224fdfc71a9a402199b3f60e32567ae5 |
| institution | Directory Open Access Journal |
| issn | 2589-0042 |
| language | English |
| last_indexed | 2024-03-11T17:50:42Z |
| publishDate | 2023-11-01 |
| publisher | Elsevier |
| record_format | Article |
| series | iScience |
| spelling | doaj.art-224fdfc71a9a402199b3f60e32567ae52023-10-18T04:31:25ZengElsevieriScience2589-00422023-11-012611108006Targeted deep learning classification and feature extraction for clinical diagnosisYiting Tsai0Vikash Nanthakumar1Saeed Mohammadi2Susan A. Baldwin3Bhushan Gopaluni4Fei Geng5University of British Columbia, 2360 East Mall, Vancouver, BC V6T 1Z3, Canada; Corresponding authorMcMaster University, 1280 Main St W, Hamilton, ON L8S 4L8, CanadaMcMaster University, 1280 Main St W, Hamilton, ON L8S 4L8, CanadaUniversity of British Columbia, 2360 East Mall, Vancouver, BC V6T 1Z3, CanadaUniversity of British Columbia, 2360 East Mall, Vancouver, BC V6T 1Z3, CanadaUniversity of British Columbia, 2360 East Mall, Vancouver, BC V6T 1Z3, CanadaSummary: Protein biomarkers can be used to characterize symptom classes, which describe the metabolic or immunodeficient state of patients during the progression of a specific disease. Recent literature has shown that machine learning methods can complement traditional clinical methods in identifying biomarkers. However, many machine learning frameworks only apply narrowly to a specific archetype or subset of diseases. In this paper, we propose a feature extractor which can discover protein biomarkers for a wide variety of classification problems. The feature extractor uses a special type of deep learning model, which discovers a latent space that allows for optimal class separation and enhanced class cluster identity. The extracted biomarkers can then be used to train highly accurate supervised learning models. We apply our methods to a dataset involving COVID-19 patients and another involving scleroderma patients, to demonstrate improved class separation and reduced false discovery rates compared to results obtained using traditional models.http://www.sciencedirect.com/science/article/pii/S2589004223020837Health sciencesComputer-aided diagnosis methodArtificial intelligence applications |
| spellingShingle | Yiting Tsai Vikash Nanthakumar Saeed Mohammadi Susan A. Baldwin Bhushan Gopaluni Fei Geng Targeted deep learning classification and feature extraction for clinical diagnosis iScience Health sciences Computer-aided diagnosis method Artificial intelligence applications |
| title | Targeted deep learning classification and feature extraction for clinical diagnosis |
| title_full | Targeted deep learning classification and feature extraction for clinical diagnosis |
| title_fullStr | Targeted deep learning classification and feature extraction for clinical diagnosis |
| title_full_unstemmed | Targeted deep learning classification and feature extraction for clinical diagnosis |
| title_short | Targeted deep learning classification and feature extraction for clinical diagnosis |
| title_sort | targeted deep learning classification and feature extraction for clinical diagnosis |
| topic | Health sciences Computer-aided diagnosis method Artificial intelligence applications |
| url | http://www.sciencedirect.com/science/article/pii/S2589004223020837 |
| work_keys_str_mv | AT yitingtsai targeteddeeplearningclassificationandfeatureextractionforclinicaldiagnosis AT vikashnanthakumar targeteddeeplearningclassificationandfeatureextractionforclinicaldiagnosis AT saeedmohammadi targeteddeeplearningclassificationandfeatureextractionforclinicaldiagnosis AT susanabaldwin targeteddeeplearningclassificationandfeatureextractionforclinicaldiagnosis AT bhushangopaluni targeteddeeplearningclassificationandfeatureextractionforclinicaldiagnosis AT feigeng targeteddeeplearningclassificationandfeatureextractionforclinicaldiagnosis |