Method for constructing an online prediction website based on biological sequence cis-acting regulatory elements

Abstract

This invention discloses a method for constructing an online prediction website based on cis-regulatory elements of biological sequences. This method is applicable to building online prediction websites from models trained using machine learning on biological sequences. With the continuous development of artificial intelligence technology, machine learning algorithms have flourished and been successfully applied in multiple fields. In particular, in the fields of biological and medical research, researchers have constructed numerous prediction models to serve various research tasks, such as the prediction of cis-regulatory elements, biomedical pathogenic elements, and proteins. Although many prediction models have been successfully constructed, professionals who frequently use computers are more concerned with how to apply these models to practical work. Therefore, constructing the prediction models in this paper into user-friendly and easy-to-use online websites can greatly enhance the practical application value of the models.

Description

Technical Field

[0001] This invention relates to the field of data processing and analysis in bioinformatics, specifically a method for constructing a general online prediction website for cis-regulatory elements in biological sequences. Background Technology

[0002] With the continuous development of artificial intelligence technology, more and more machine learning algorithms and models have been proposed and successfully applied to multiple fields. In particular, given the current trend of highly interdisciplinary integration, especially in bioinformatics, researchers have built a large number of predictive models to advance various research questions. For example, in biological research, scientists annotate genomic data to improve gene structural information or predict specific proteins to explore the pathogenic mechanisms of bacteria, viruses, etc. Although a large number of predictive models have been built, they rarely demonstrate practical application value because these academic models are difficult for ordinary practitioners to apply; on the other hand, with the increasing integration of computers into people's lives and work, numerous websites have been built to meet people's various needs.

[0003] Therefore, building a user-friendly and easy-to-use online prediction website can maximize the practicality of the paper's model. Summary of the Invention

[0004] In view of the problems pointed out in the prior art, the present invention provides a method for constructing an online prediction website based on cis-regulatory elements of biological sequences. This method targets biological sequences and utilizes multiple website design techniques to improve the usability of the paper model.

[0005] The technical solution to achieve the objective of this invention is:

[0006] Methods for constructing online prediction websites based on biological sequence cis-regulatory elements include:

[0007] (1) Data preprocessing: Preprocess the input data file according to the model;

[0008] (2) Construct the main presentation layer of the website: Design the UI of each front-end page according to the specific application scenario and user prototype of the website;

[0009] (3) Construct the main service interaction layer of the website: Data interaction is performed between the Java data terminal and the front-end page, and the user-submitted request task is passed to the Perl computing terminal. Using the Struts2 framework, the front-end page constructed in step (2) is connected to the Java data terminal, and the preprocessed data is encapsulated into JSON format and passed to the Perl computing terminal. At the same time, the Hibernate framework is used to map the DAO layer variables of the Java data terminal to the database fields, encapsulate the active connection of the Java database, convert the data model into a storage model, and perform data persistence.

[0010] (4) Construct the main business logic layer of the website: After processing the data submitted by the front-end page through the Java data terminal, the Perl script is called to predict the cis-acting control elements using the machine learning model; the distributed task distribution framework Gearman is used for task scheduling. When the request task is received from the Java data terminal in step (3), Gearman adds the data to the task pool and executes it in sequence; when the number of submitted request tasks exceeds the Worker value, the extra tasks are placed in the waiting area until a task in execution ends before calculation begins.

[0011] (5) Construct the main persistent layer of the website: When the business logic layer in step (4) completes the calculation task, the prediction results are written into the MySQL database to prepare for the output of the results;

[0012] (6) Output prediction results: The Java data terminal retrieves data from the database and displays it on the front-end page.

[0013] Furthermore, the data files in step (1) include two formats, fasta and fastq, for storing DNA, RNA and protein sequences.

[0014] Furthermore, the UI design of the front-end page in step (2) includes the overall style of the page, page elements, operation controls, and user guidance.

[0015] The beneficial effects of this invention are: by processing common sequencing data files and using mainstream B / S (browser and server) architecture technology, models trained by traditional machine learning or deep learning can be built into user-friendly and easy-to-use online prediction websites, improving the usability of paper models and thus serving the research work of practitioners in related fields. Attached Figure Description

[0016] Figure 1 This is a model diagram for constructing an online prediction website according to an embodiment of the present invention. Detailed Implementation

[0017] The present invention will be further described below with reference to the accompanying drawings and embodiments, but this is not intended to limit the scope of the invention.

[0018] Example:

[0019] A general method for constructing an online prediction website for biological sequence cis-regulatory elements includes the following three parts:

[0020] 1) Data Preprocessing: Common sequencing file formats include FASTA and FASTQ, the difference being that FASTQ offers additional sequencing quality control. In this section, necessary preprocessing is required based on the model's requirements for the input data. For example, for DNA and RNA sequences, a 4-bit one-hot encoding method can be used, encoding a sequence of length N into an N×4 two-dimensional matrix; for protein sequences, a suitable encoding method may need to be selected.

[0021] 2) The website's main structure is divided into a presentation layer, a service interaction layer, a business logic layer, and a persistence layer, such as... Figure 1 As shown. The construction of each layer includes the following steps:

[0022] Presentation layer:

[0023] 2-1): Based on the specific application scenario and user prototype of the website, design the UI of each front-end page, including the overall style, page elements, operation controls, and user guidance. This step requires the use of multiple technical frameworks such as jQuery, Bootstrap, Ajax, and Google Charts. For example, for an RNA splicing site prediction website, its front-end page may include a data import button (FASTA file) or a long text box (sequence fragment), and may also include a selection list (multiple species and donor / recipient) due to the need for multi-species prediction models, as well as numerical input boxes depending on whether the user can customize the prediction threshold. In addition, pages for model introduction, user help, and data download can be designed to improve the usability of the website;

[0024] Service interaction layer:

[0025] 2-2): The Java data client is responsible for interacting with the front-end page and passing user-submitted requests to the Perl computing client. This step uses the Struts2 framework to connect the front-end page from step 2-1) with the Java data client, encapsulating the pre-processed data into JSON format and passing it to the Perl computing client. Furthermore, the Hibernate framework is used to map the DAO (Data Access Object) layer variables of the Java data client to database fields, encapsulating the active connection to the Java database for easy implementation of CRUD operations. Based on the Hibernate framework, the MySQL database can convert the data model into a storage model for data persistence. Taking the Linux Ubuntu operating system as an example, the specific steps for setting up the environment are as follows:

[0026] S01: Java Installation

[0027] Download a compatible version of the JDK (Java Development Kit) from the Oracle website, extract it to the / usr / lib / jvm / directory on your system, and modify the ~ / .bashrc file to configure the Java environment variables.

[0028] S02: Tomcat Installation

[0029] Tomcat is a free web server software. It can be installed using the command `sudo apt-get install tomcatX`, where X represents the Tomcat version number. After installation, modify the memory allocation in the configuration file according to the expected website traffic.

[0030] S03: Installation of PHP and Apache

[0031] The installation order must be PHP first, then Apache. Execute the commands `sudo apt-get install phpX` and `sudo apt-get install apacheY` to install PHP and Apache respectively, where X and Y represent the version numbers of PHP and Apache, respectively.

[0032] S04: Configure Apache Reverse Proxy

[0033] Execute the command `sudo vim / etc / apache2 / apache2.conf` to modify the redirection information, and then execute the command `sudo a2enmod proxy_http` to ensure that the changes take effect;

[0034] S05: Installation of MySQL Database

[0035] Execute the command `sudo apt-get install mysql-server` to install, and modify the database login password as needed;

[0036] S06: Installing phpMyAdmin

[0037] Execute the command `sudo apt-get install phpmyadmin` to install it;

[0038] Business logic layer:

[0039] 2-3): After processing the data submitted by the front-end page, the Java side calls a Perl script to use a machine learning model to predict cis-regulatory components. This step is typically resource- and time-intensive. To ensure server load balancing when multiple tasks are submitted, this layer uses the distributed task distribution framework Gearman for task scheduling. Upon receiving the request tasks from the Java side in step 2-2), Gearman adds them to the task pool and executes them sequentially. Furthermore, the server can set the number of worker processes, i.e., the number of computational tasks processed simultaneously. When the number of submitted request tasks exceeds the worker value, the extra tasks enter a waiting area and only begin computation after one of the currently executing tasks finishes. Taking the Linux Ubuntu operating system as an example, the specific steps for setting up the environment are as follows:

[0040] S01: Configure Apache CGI

[0041] Execute the command `sudo cpan` to open cpan (Comprehensive Perl Archive Network), then execute `install CGI` to install the CGI (Common Gateway Interface) module. Next, modify the root directory of the CGI program by executing the command `sudo vim / etc / apache2 / apache2.conf`, and then perform a page test.

[0042] S02: Installation of Perl related modules

[0043] Install three modules under CPAN:

[0044] install Storable

[0045] Install DBI

[0046] install IO::All

[0047] Execute command:

[0048] sudo apt-get install libgd-dev

[0049] sudo apt-get install libxml-xpath-perl

[0050] Install eight modules under CPAN:

[0051] install Bio::SeqIO

[0052] install Bio::Seq

[0053] install Capture::Tiny

[0054] install SOAP::Transport::HTTP

[0055] install MIME::Lite

[0056] install Mail::Sendmail

[0057] Install Email::MIME

[0058] install Email::Sender::Simple

[0059] S03: Gearman Message Queue Installation

[0060] First, install the following Gearman dependencies:

[0061] sudo apt-get install libboost-dev

[0062] sudo apt-get install libboost-all-dev

[0063] sudo apt-get install gperf

[0064] sudo apt-get install libevent-dev

[0065] sudo apt-get install uuid-dev

[0066] sudo apt-get install libssl-dev

[0067] Download and install Gearman:

[0068] cd ~

[0069] wget https: / / launchpad.net / gearmand / 1.2 / 1.1.12 / +download / gearmand-1.1.12.tar.gz

[0070] tar zxvf gearmand-1.1.12.tar.gz

[0071] cd gearmand-1.1.12 /

[0072] sudo make clean

[0073] . / configure

[0074] sudo make

[0075] sudo make install

[0076] Download and install Gearman's job server:

[0077] sudo apt-get install gearman-job-server

[0078] Install three modules under CPAN:

[0079] install Net::SSLeay

[0080] install IO::Socket::SSL

[0081] install Gearman::Util

[0082] S04: Install the Perl module for connecting to MySQL

[0083] sudo cpan

[0084] install DBD::mysql

[0085] S05: Python Installation

[0086] sudo apt-get install python-scipy

[0087] sudo apt-get install python-dev

[0088] sudo apt-get install python-numpy

[0089] sudo apt-get install python-matplotlib

[0090] sudo apt-get install python-pandas

[0091] sudo apt-get install python-sklearn

[0092] S06: R Language Installation

[0093] sudo apt-get install r-base

[0094] sudo R

[0095] Persistent layer:

[0096] 2-4): When the calculation task in step 2-3) of the business logic layer is completed, the prediction result is written to the MySQL database to prepare for the output of the result;

[0097] 3) Prediction Result Output: The Java database retrieves data from the database and displays it on the front end. For example, for an RNA splicing site prediction website, if the model is used to predict classic splicing sites (donor GT, recipient AG), its output should provide the predicted values ​​of characteristic sites within the effective sequence region, which users can filter according to a custom threshold.

[0098] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the concept described herein through the above teachings or related technologies or knowledge. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

Claims

1. A method for constructing an online prediction website based on cis-regulatory elements of biological sequences, characterized by: include: (1) Data preprocessing: The input data file is preprocessed according to the model; the data file includes two formats, fasta and fastq, for storing DNA, RNA and protein sequences; (2) Construct the main presentation layer of the website: Design the UI of each front-end page according to the specific application scenario and user prototype of the website; (3) Construct the main service interaction layer of the website: Data interaction is performed between the Java data terminal and the front-end page, and the user-submitted request task is passed to the Perl computing terminal. Using the Struts2 framework, the front-end page constructed in step (2) is connected to the Java data terminal, and the preprocessed data is encapsulated into JSON format and passed to the Perl computing terminal. At the same time, the Hibernate framework is used to map the DAO layer variables of the Java data terminal to the database fields, encapsulate the active connection of the Java database, convert the data model into a storage model, and perform data persistence. (4) Construct the main business logic layer of the website: After processing the data submitted by the front-end page through the Java data terminal, call the Perl script and use the machine learning model to predict the cis-acting regulatory elements; The distributed task distribution framework Gearman is used for task scheduling. When the request task is received from the Java data terminal in step (3), Gearman adds the data to the task pool and executes it in sequence. When the number of submitted request tasks exceeds the Worker value, the extra tasks are placed in the waiting area until a task in execution finishes before calculation begins. (5) Construct the main persistent layer of the website: When the business logic layer in step (4) completes the calculation task, the prediction results are written into the MySQL database to prepare for the output of the results; (6) Output prediction results: The Java data terminal retrieves data from the database and displays it on the front-end page.

2. The online prediction website construction method according to claim 1, characterized in that: The UI design of the front-end page in step (2) includes the overall style of the page, page elements, operation controls and user guidance.

Images