Exploring the potential of Artificial Immunologic Sistems in Chromatin structural prediction pipeline

• Grant number: 25/25221-7
• Support Opportunities: Scholarships abroad - Research Internship - Scientific Initiation
• Start date: July 06, 2026
• End date: October 06, 2026
• Field of knowledge: Engineering - Biomedical Engineering - Bioengineering
• Principal Investigator: Fábio Roberto Chavarette
• Grantee: Gustavo Soares
• Supervisor: Vinicius de Godoi Contessoto
• Host Institution: Instituto de Química (IQ). Universidade Estadual Paulista (UNESP). Campus de Araraquara. Araraquara , SP, Brazil
• Institution abroad: Rice University, United States
• Associated to the scholarship: 25/08602-7 - Development of an Intelligent Tool Oriented to Analysis of Electrophoresis Band Images, BP.IC

Abstract

The organization of the chromosome is an area that deeply intrigues a vast portionof academia, and for good reason. Studies over time have shown the functional dependencebetween a tissue's cellular expression profile and the structural conformation of chromatin,precisely because the accessibility to the information contained in the chromosome isbased on more or less coiled chromatin regions. This relationship was also observedand correlated with different cancer profiles, where topologically associating domains(TADs), laminar-associated domains, and even interactions with promoters exhibitedanomalous characteristics observable in cancerous cells at initial and more advanced stages,giving them their typical profiles in terms of behavior and effects on the organism. Inthe search for understanding these phenomena, specific spectroscopy and sequencingtechniques for the structural determination of the chromosome have been developed. Themost well-known among these are Fluorescence In Situ Hybridization (FISH) and HighthroughputChromosome Conformation Capture (HI-C), in addition to techniques seekingfunctional understanding of the chromosome using epigenetic marks, with ChromatinImmunoprecipitation followed by Sequencing (ChIP-Seq) being one of the most notable.With the data obtained from these techniques (mainly from HI-C assays), and with the aidof computational tools, it was possible to establish models to predict the three-dimensionalstructure of the chromosome. However, various problems were observed regarding thedifficulty of obtaining de novo data, related to the effort required to acquire high-resolutionchromosome contact frequency maps, coupled with the significant endeavor necessary forHI-C assays. As a result, alternatives utilizing ChIP-Seq data and machine learning havebeen proposed, highlighting solutions that use artificial neural networks oriented towardsphysical models, such as PyMEGABASE and TECSAS.Nevertheless, it is still necessaryto evaluate alternatives seeking better results, or to complement those already obtained.Thus, Artificial Immune Systems (AIS), specifically Negative Selection Algorithms (NSA),present themselves as a potential avenue to be employed both in identifying and correctingnoise, and in performing tasks like PyMEGABASE and TECSAS in the chromosomeanalysis pipeline. Therefore, the objective of this study is to evaluate these two possibilitiescomparatively, thereby contributing to the field with the insertion of AIS into a new areaof study.