Classification of hemoglobin fractions in the liquid state using Raman spectroscopy combined with machine learning

Sara Abbasi; Mehdi Feizpour; Ilse Weets; Qing Liu; Hugo Thienpont; Francesco Ferranti; Heidi Ottevaere

doi:10.1016/j.microc.2023.109305

Classification of hemoglobin fractions in the liquid state using Raman spectroscopy combined with machine learning

Sara Abbasi, Mehdi Feizpour, Ilse Weets, Qing Liu, Hugo Thienpont, Francesco Ferranti, Heidi Ottevaere

Research output: Contribution to journal › Article › peer-review

7 Citations (Scopus)

91 Downloads (Pure)

Abstract

Identifying hemoglobinopathies is important for the clinical management of many diseases. One of the common techniques to screen hemoglobinopathies is through high-performance liquid chromatography separation followed by UV–VIS detection. Although UV–VIS can quantify the hemoglobin fractions, it is unable to identify them. Here, we use Raman microscopy to generate fingerprint spectra of hemoglobin fractions based on which the fractions can be identified. Five different hemoglobin types are investigated in their liquid state: HbA0, HbS, HbF, HbA1c, and HbA2. Machine learning models based on support vector machines and fully-connected neural networks are optimized to classify these fractions achieving 98.2 ± 0.1% and 98.5 ± 0.3% test F1-score, respectively. In addition, the test accuracy of these two models are 98.2 ± 0.1% and 98.5 ± 0.3%, respectively. Our approach demonstrates the potential of Raman spectroscopy as an identification module in combination with high-performance liquid chromatography. Moreover, this detection approach can be easily miniaturized and integrated with microfluidics.

Original language	English
Article number	109305
Number of pages	7
Journal	Microchemical Journal
Volume	194
DOIs	https://doi.org/10.1016/j.microc.2023.109305
Publication status	Published - Nov 2023

Bibliographical note

Funding Information:
The authors acknowledge funding by the European Union within the Horizon Europe Program, under the EIC Pathfinder Project “VortexLC”, Grant Agreement no. 101047029 and by Brussels Hoofdstedelijk Gewest - Innoviris, under the VortexDiaLC project (2022-RPF-2). This work was also supported in part by the Methusalem and Hercules foundations and the OZR of the Vrije Universiteit Brussel (VUB).

Publisher Copyright:
© 2023 The Author(s)

Keywords

Hemoglobin variantsBiosensorsRaman spectroscopyPoint-of-care testingHigh-Performance Liquid Chromatography (HPLC)MicrofluidicsClinical managementMachine learning optimization

Access to Document

10.1016/j.microc.2023.109305Licence: CC BY-NC-ND

98840524Final published version, 3.05 MBLicence: CC BY-NC-ND

Cite this

@article{9770e3ab422e4b49aa0a49b6a58e7035,

title = "Classification of hemoglobin fractions in the liquid state using Raman spectroscopy combined with machine learning",

abstract = "Identifying hemoglobinopathies is important for the clinical management of many diseases. One of the common techniques to screen hemoglobinopathies is through high-performance liquid chromatography separation followed by UV–VIS detection. Although UV–VIS can quantify the hemoglobin fractions, it is unable to identify them. Here, we use Raman microscopy to generate fingerprint spectra of hemoglobin fractions based on which the fractions can be identified. Five different hemoglobin types are investigated in their liquid state: HbA0, HbS, HbF, HbA1c, and HbA2. Machine learning models based on support vector machines and fully-connected neural networks are optimized to classify these fractions achieving 98.2 ± 0.1% and 98.5 ± 0.3% test F1-score, respectively. In addition, the test accuracy of these two models are 98.2 ± 0.1% and 98.5 ± 0.3%, respectively. Our approach demonstrates the potential of Raman spectroscopy as an identification module in combination with high-performance liquid chromatography. Moreover, this detection approach can be easily miniaturized and integrated with microfluidics.",

keywords = "Hemoglobin variantsBiosensorsRaman spectroscopyPoint-of-care testingHigh-Performance Liquid Chromatography (HPLC)MicrofluidicsClinical managementMachine learning optimization",

author = "Sara Abbasi and Mehdi Feizpour and Ilse Weets and Qing Liu and Hugo Thienpont and Francesco Ferranti and Heidi Ottevaere",

note = "Funding Information: The authors acknowledge funding by the European Union within the Horizon Europe Program, under the EIC Pathfinder Project “VortexLC”, Grant Agreement no. 101047029 and by Brussels Hoofdstedelijk Gewest - Innoviris, under the VortexDiaLC project (2022-RPF-2). This work was also supported in part by the Methusalem and Hercules foundations and the OZR of the Vrije Universiteit Brussel (VUB). Publisher Copyright: {\textcopyright} 2023 The Author(s)",

year = "2023",

month = nov,

doi = "10.1016/j.microc.2023.109305",

language = "English",

volume = "194",

journal = "Microchemical Journal",

issn = "0026-265X",

publisher = "Elsevier Inc.",

}

TY - JOUR

T1 - Classification of hemoglobin fractions in the liquid state using Raman spectroscopy combined with machine learning

AU - Abbasi, Sara

AU - Feizpour, Mehdi

AU - Weets, Ilse

AU - Liu, Qing

AU - Thienpont, Hugo

AU - Ferranti, Francesco

AU - Ottevaere, Heidi

N1 - Funding Information: The authors acknowledge funding by the European Union within the Horizon Europe Program, under the EIC Pathfinder Project “VortexLC”, Grant Agreement no. 101047029 and by Brussels Hoofdstedelijk Gewest - Innoviris, under the VortexDiaLC project (2022-RPF-2). This work was also supported in part by the Methusalem and Hercules foundations and the OZR of the Vrije Universiteit Brussel (VUB). Publisher Copyright: © 2023 The Author(s)

PY - 2023/11

Y1 - 2023/11

N2 - Identifying hemoglobinopathies is important for the clinical management of many diseases. One of the common techniques to screen hemoglobinopathies is through high-performance liquid chromatography separation followed by UV–VIS detection. Although UV–VIS can quantify the hemoglobin fractions, it is unable to identify them. Here, we use Raman microscopy to generate fingerprint spectra of hemoglobin fractions based on which the fractions can be identified. Five different hemoglobin types are investigated in their liquid state: HbA0, HbS, HbF, HbA1c, and HbA2. Machine learning models based on support vector machines and fully-connected neural networks are optimized to classify these fractions achieving 98.2 ± 0.1% and 98.5 ± 0.3% test F1-score, respectively. In addition, the test accuracy of these two models are 98.2 ± 0.1% and 98.5 ± 0.3%, respectively. Our approach demonstrates the potential of Raman spectroscopy as an identification module in combination with high-performance liquid chromatography. Moreover, this detection approach can be easily miniaturized and integrated with microfluidics.

AB - Identifying hemoglobinopathies is important for the clinical management of many diseases. One of the common techniques to screen hemoglobinopathies is through high-performance liquid chromatography separation followed by UV–VIS detection. Although UV–VIS can quantify the hemoglobin fractions, it is unable to identify them. Here, we use Raman microscopy to generate fingerprint spectra of hemoglobin fractions based on which the fractions can be identified. Five different hemoglobin types are investigated in their liquid state: HbA0, HbS, HbF, HbA1c, and HbA2. Machine learning models based on support vector machines and fully-connected neural networks are optimized to classify these fractions achieving 98.2 ± 0.1% and 98.5 ± 0.3% test F1-score, respectively. In addition, the test accuracy of these two models are 98.2 ± 0.1% and 98.5 ± 0.3%, respectively. Our approach demonstrates the potential of Raman spectroscopy as an identification module in combination with high-performance liquid chromatography. Moreover, this detection approach can be easily miniaturized and integrated with microfluidics.

KW - Hemoglobin variantsBiosensorsRaman spectroscopyPoint-of-care testingHigh-Performance Liquid Chromatography (HPLC)MicrofluidicsClinical managementMachine learning optimization

UR - http://www.scopus.com/inward/record.url?scp=85170428746&partnerID=8YFLogxK

U2 - 10.1016/j.microc.2023.109305

DO - 10.1016/j.microc.2023.109305

M3 - Article

VL - 194

JO - Microchemical Journal

JF - Microchemical Journal

SN - 0026-265X

M1 - 109305

ER -

Classification of hemoglobin fractions in the liquid state using Raman spectroscopy combined with machine learning

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this