Highly Performing Automatic Detection of Structural Chromosomal Abnormalities Using Siamese Architecture. Issue 8 (15th April 2023)

• Record Type: Journal Article
• Authors: Bechar, Mohammed El Amine; Guyader, Jean-Marie; El Bouz, Marwa; Douet-Guilbert, Nathalie; Al Falou, Ayman; Troadec, Marie-Bérengère
• Is Part Of: Journal of molecular biology. Volume 435:Issue 8(2023)
• Publication Date: 2023-04-15
• Subjects: cytogenetics -- structural chromosomal abnormalities -- convolutional neural networks -- siamese architecture -- deletion/inversion detection; Molecular biology -- Periodicals; Biology -- Periodicals; Biochemistry -- Periodicals; Bacteriology -- Periodicals; Molecular Biology -- Periodicals; Biochemistry -- Periodicals; Biologie moléculaire -- Périodiques; Biologie -- Périodiques; Biochimie -- Périodiques; Moleculaire biologie; Biochemistry; Biology; Molecular biology; Periodicals; 572.805
• DOI: 10.1016/j.jmb.2023.168045
• Languages: English
• ISSNs: 0022-2836

Abstract

Graphical abstract: Highlights: We describe a novel technique for detecting structural chromosomal abnormalities. The Siamese architecture that we use mimics how human operators detect abnormalities. This machine learning method takes as inputs both copies of each chromosome pair. The method detects changes in the banding pattern of the chromosomes of a given pair. We identify adapted margins of the contrastive loss function for the Siamese training. Abstract: The detection of structural chromosomal abnormalities (SCA) is crucial for diagnosis, prognosis and management of many genetic diseases and cancers. This detection, done by highly qualified medical experts, is tedious and time-consuming. We propose a highly performing and intelligent method to assist cytogeneticists to screen for SCA. Each chromosome is present in two copies that make up a pair of chromosomes. Usually, SCA are present in only one copy of the pair. Convolutional neural networks (CNN) with Siamese architecture are particularly relevant for evaluating similarities between two images, which is why we used this method to detect abnormalities between both chromosomes of a given pair. As a proof-of-concept, we first focused on a deletion occurring on chromosome 5 (del(5q)) observed in hematological malignancies. Using our dataset, we conducted several experiments without and with data augmentation on seven popular CNN models. Overall, performances obtained were very relevant for detecting deletions, particularly with Xception and InceptionResNetV2 models achieving 97.50 % and 97.01 % of F 1 -score, respectively. We additionally demonstrated that these models successfully recognized another SCA, inversion inv(3), which is one of the most difficult SCA to detect. The performance improved when the training was applied on inversion inv(3) dataset, achieving 94.82 % of F 1 -score. The technique that we propose in this paper is the first highly performing method based on Siamese architecture that allows the detection of SCA. Our code is publicly available at: https://github.com/MEABECHAR/ChromosomeSiameseAD .