Providing a suitable data model to discover the transmission of genetic diseases

Computer Engineering Master Thesis

Software Orientation

Abstract

In today's society, due to the progress of medical science, the volume of medical data is increasing rapidly. For faster and more efficient analysis of these data, we need electronic storage of these data. The data related to genetic diseases are also considered from this category of data. Considering this issue, we should design suitable databases for storing and retrieving these data. Considering the nature of genetic data and the problem of transmission of genetic diseases, relationships between people and its analysis is considered an important issue, in this article we use the graph data model, which is one of the subsets of the non-structured data model (NOSQL), to store and retrieve this data. For this purpose, we first determine the needs and queries related to this issue and based on that, we design the data graphic model. To evaluate the designed data model, a team of genetic experts also reviewed this data model and expressed their favorable opinion regarding the use of this data model for genetic diseases. We also used the Neo4j software, which stored the data related to the genetic disease Thalassemia, and we examined this data model based on the efficiency of storing and retrieving information and the time of the queries. Considering the time of the queries and the lack of support of other data models for the relationships between people, this data model is considered a suitable model.

Key words: genes, genetic diseases, graph databases, neo4j, data model

Chapter First: Introduction

1-1-Preface

In the field of medicine, data are being produced and spread rapidly. These data are produced in different forms than the data of the past, and according to the advances of science in this field, the need for new management is felt much more than in the past. To store these data, we need databases that are capable of supporting various types of data and a large volume of data, and also have the ability to perform correct and complete data management [14].

Regarding genetic diseases, the data that we need to store are diverse data. According to the nature of genetic diseases, in order to understand how these diseases are transmitted, we also need to save the health status of the ancestors of the patients. In each investigation, a new person may be added to this genealogy. Also, to discover the path of transmission of diseases, the relationships between people in this database are very important and necessary. To support these needs and manage relationships between people and disease transmission, structured databases [1] are not a suitable option, because they are not able to support different types of data.

Unstructured databases [2] are more suitable options to support different types of data. There are different types of NOSQL databases, but due to the fact that in this type of disease, relationships between people are very important and we must be able to add entities at any time, graph databases are the right case. The cell nucleus consists of 46 chromosomes or 23 pairs of chromosomes. Chromosomes have tangled strands called DNA, which contain genes. Each cell of the human body contains 25,000 to 35,000 genes [1]. Genes contain information that makes up human characteristics. Genes are composed of pairs of bases called nucleotides. The basic composition consists of 4 structures: Adenine [3] / Guanine [4] / Cytosine [5] / Thymine [6] In fact, we use 4 letters A, T, C, G to express each gene, which is called the nucleotide sequence. The nucleotide sequences of different diseases are different from each other. For example, the length of the insulin gene sequence is 333 characters. So far, the longest known nucleotide sequence is related to Duchenne's disease. The length of this sequence is 2.3 megabytes. As examples of genetic diseases, leprosy, skin cancers, mental retardation, sickle cell anemia, phenylketonuria, thalassemia and the like can be mentioned [2].

Some of the cases in which genetic tests are performed are as follows:

• A couple who is planning to start a family and one of them or one of his close relatives has a hereditary disease.

• A person who has a child with an acute birth defect.

• A child with a physical problem that can be genetic.

In order to perform these genetic tests in the first place, you need to have We have the genealogy of couples when forming a family, parents during pregnancy tests, and patients when examining a genetic patient. After knowing the genealogy, when dealing with diseases, we need to store information about the patient. To store data related to genetic diseases, we need a database that can well support the storage of all types of data. For this data, we need a data model that can analyze and analyze this data in addition to storing it. One of the important issues for choosing a data model is that in the case of medical data for each patient, we may save characteristics that we do not need for other patients, for example, we may need to save blood test results for one patient, but we do not need to perform this test and save this test for another patient, or we may come across things that were not foreseen during the examination of the patient's condition, for this reason, it is better not to design a general plan for the database from the beginning in order to be able to save any characteristic that we needed or while working with We encountered it and added it for the patient. Based on this issue, we come to the conclusion that we cannot use SQL databases and NOSQL databases are more suitable for this. Another important issue is that due to the need for the genetics of family members and previous generations of the patient, we must have the ability to add new entities (previous and subsequent generations) to this database while conducting research. Regarding the transmission of diseases, it is very important to determine the route of transmission of the disease because it must be determined that the disease was inherited from the father or mother, which of the ancestors in the next stage, and also to determine which of the male or female children may inherit the disease. For this reason, we must design databases that have the ability to extract relationships between entities. Of course, the relationships between the entities in the interface databases can also be extracted, but firstly, it is very complicated and time-consuming due to the need to write nested procedures. Considering these three issues, we come to the conclusion that the graph database is the best choice for this type of disease.

       In this thesis, we will design a database using the graph data model that has the ability to store different types and volumes of data. The designed database must have the ability to perform operations on these stored data and be able to extract the desired results from it in investigating the transmission of genetic diseases. Results such as the path of disease transmission, the possibility of disease transmission to the next generation or the possibility of disease transmission to a certain gender of the next generation, the percentage of disease transmission, etc. In this database, the entities that are the same people are stored in the nodes, in addition to the general characteristics of the patients, all the information related to the diseases of the people, conditions and symptoms of the patients are also stored in the nodes. In the next levels of the graph, higher generations of patients will be stored along with information about the specific disease they are researching. We will use

edges to display the relationships between people in this data model. In this way, if the disease is transmitted from one person to another, we will use the directed edge to show this transmission. In addition to these, we can also add explanations on the ridges. Explanations such as the percentage of the probability of transmission of a specific disease from one person to another. 1-3-Importance and necessity of conducting research So far, several data models have been used to store medical information, but each of these data models has disadvantages that make them not ideal data models. One of these data models is the interface data model.