The work carried out during the Ph.D. in Medical Sciences and Biotechnology course tested the application of machine learning models in several precision medicine topics. In bioinformatic biomarker analysis, the publications focused on discretizing the gene expression levels to obtain a manageable and insightful granularity. The works demonstrated novel analysis pipelines to detect survival and tumor stages from oncologic patients’ biomarkers. The same chapter presented a procedure for a public health decision support system based on machine learning, which has also been demonstrated on the same dataset. The chemoinformatics numerical experiments for drug toxicity, bioaccumulation prediction, or P450 enzyme bioactivity evaluation all exploited spiking neural networks, showing the ability of this technique to handle structural information of the compounds for predictive analysis. For clinical precision medicine, an algorithm has been tested fusing clinical variables (ordinal and binary) from nearly 300 patients to forecast the risk of developing lymphedema after breast cancer therapy. Moreover, free software has been released to measure the volumetry of the affected limb in case of edema or other pathologies requiring tracking of body parts over time. Another chapter reported the development of a free Python library to run equivalence tests in the biomedical sector, focusing on advanced visualization of the statistical outcomes. This library also fills a gap in the biostatistical tools available to Python users requiring biomedical equivalence analysis. Regarding regenerative medicine, a study has been introduced to track octacalcium phosphate synthesis through a machine-learning methodology centered on a novel algorithm exploiting an ad-hoc solution on merged XRD and FTIR peak descriptors. Octacalcium phosphate is found in biological systems, particularly in the early bone formation and mineralization stages. It is a precursor to hydroxyapatite, the main mineral component of bones and teeth. The last chapter introduced a mass spectrometry proteomic analysis sequence to detect aberrant protein expression levels. The procedure has been tested on mesenchymal stem cells’ extracellular vesicle protein content cultured on biomaterials doped or not with metallic ions.

Machine learning approaches for personalized medicine / Nascimben, Mauro. - ELETTRONICO. - (2024 Feb 22).

Machine learning approaches for personalized medicine

Nascimben, Mauro
2024-02-22

Abstract

The work carried out during the Ph.D. in Medical Sciences and Biotechnology course tested the application of machine learning models in several precision medicine topics. In bioinformatic biomarker analysis, the publications focused on discretizing the gene expression levels to obtain a manageable and insightful granularity. The works demonstrated novel analysis pipelines to detect survival and tumor stages from oncologic patients’ biomarkers. The same chapter presented a procedure for a public health decision support system based on machine learning, which has also been demonstrated on the same dataset. The chemoinformatics numerical experiments for drug toxicity, bioaccumulation prediction, or P450 enzyme bioactivity evaluation all exploited spiking neural networks, showing the ability of this technique to handle structural information of the compounds for predictive analysis. For clinical precision medicine, an algorithm has been tested fusing clinical variables (ordinal and binary) from nearly 300 patients to forecast the risk of developing lymphedema after breast cancer therapy. Moreover, free software has been released to measure the volumetry of the affected limb in case of edema or other pathologies requiring tracking of body parts over time. Another chapter reported the development of a free Python library to run equivalence tests in the biomedical sector, focusing on advanced visualization of the statistical outcomes. This library also fills a gap in the biostatistical tools available to Python users requiring biomedical equivalence analysis. Regarding regenerative medicine, a study has been introduced to track octacalcium phosphate synthesis through a machine-learning methodology centered on a novel algorithm exploiting an ad-hoc solution on merged XRD and FTIR peak descriptors. Octacalcium phosphate is found in biological systems, particularly in the early bone formation and mineralization stages. It is a precursor to hydroxyapatite, the main mineral component of bones and teeth. The last chapter introduced a mass spectrometry proteomic analysis sequence to detect aberrant protein expression levels. The procedure has been tested on mesenchymal stem cells’ extracellular vesicle protein content cultured on biomaterials doped or not with metallic ions.
22-feb-2024
XXXVI
Chemistry and Biology
File in questo prodotto:
File Dimensione Formato  
Nascimben_tesi_2024.pdf

file ad accesso aperto

Descrizione: PDF M. Nascimben tesi di dottorato
Tipologia: Tesi di dottorato
Licenza: DRM non definito
Dimensione 9.4 MB
Formato Adobe PDF
9.4 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11579/174762
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? 2
social impact