Unlock instant, AI-driven research and patent intelligence for your innovation.

Non-invasive diagnosis method for infectious diseases based on deep learning

A deep learning and disease technology, applied in neural learning methods, medical automated diagnosis, informatics, etc., can solve problems such as insufficient extraction of representative antibody library, low prediction accuracy of disease diagnosis model, etc., to improve the accuracy of model diagnosis, The effect of improving prediction accuracy

Pending Publication Date: 2022-05-17
GUANGDONG GENERAL HOSPITAL
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] Aiming at the shortcomings of insufficient feature extraction for characterizing antibody repertoires and low prediction accuracy of disease diagnosis models in the prior art, the present invention provides a method for non-invasive diagnosis of infectious diseases based on deep learning, comprising the following steps:

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Non-invasive diagnosis method for infectious diseases based on deep learning
  • Non-invasive diagnosis method for infectious diseases based on deep learning
  • Non-invasive diagnosis method for infectious diseases based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0036] Example 1: Training set and test set selection

[0037] Collect a certain number of pathogen-infected samples and a certain number of healthy control samples, and randomly assign all collected pathogen-infected samples and healthy control samples according to a certain ratio as a training set and a test set. The number of samples of each category is not less than 100, preferably not less than 200. The ratio is 2:1˜5:1, preferably 2.5:1˜4.5:1.

Embodiment 2

[0038] Example 2: Antibody library data acquisition

[0039] (1) Sample collection

[0040] Take 1 ml of human peripheral blood samples (Peripheral Blood Mononuclear Cells, PBMCs) from the pathogen-infected patients and healthy controls in Example 1, collect them in anticoagulant tubes containing EDTA, and store them at room temperature for no more than 4 hours. PBMCs were separated by density gradient centrifugation using lymphatic separation medium (Axis-Shield, 1114547), and the separated cells were lysed in RLT buffer (Qiagen), added with 1% β-mercaptoethanol (Sigma), and then stored at -80°C For short term storage.

[0041] (2) RNA extraction, reverse transcription

[0042] RNA was extracted using RNeasy Mini Kit (Qiagen, 74106) according to the instructions, RNA concentration was measured using NanoDrop 2000c, and 500ng of each sample was taken for reverse transcription.

[0043] Reverse transcription using Thermo SuperScript TM II Reverse Transcriptase and Takara's...

Embodiment 3

[0060] Example 3: Sequencing data processing

[0061] The downstream analysis was performed on the initial sequencing data (FASTQ file) obtained in Example 2 using MiXCR (version 3.0.7) software. Analysis with the same V / J gene and CDR3 nucleic acid sequence was defined as the same antibody molecule (clone), and only IgG antibody molecules were retained. The analysis parameters are as follows:

[0062] Align: mixcr align --library my_library -t 8 -r align_log.txt R1 R2 alignments.vdjca -s hs

[0063] Assemble: mixcr assemble -r assemble_log.txt -OseparateByV=true - OseparateByJ=true -OseparateByC=true alignments.vdjcaclones.clna

[0064] Export clones: mixcrexportClones–c IGH clones.clna clones.txt

[0065] Export Alignments: mixcrexportAlignments -f -readIds -cloneId -vHit - vAlignment -jHit -jAlignment -cHit -cAlignment -nFeature FR1 -nFeature CDR1 - nFeature FR2 -nFeature CDR2 -nFeature FR3 -nFeature CDR3 -nFeature FR4 - aaFeature FR1 -aaFeature CDR1 -aaF...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a non-invasive diagnosis method and system for infectious diseases based on deep learning. The non-invasive diagnosis method is characterized by comprising the following steps: acquiring sample antibody group library data, and determining a training set and a test set; aiming at the antibody group library data, extracting group library horizontal features and sequence horizontal features; respectively constructing initial prediction models by using the extracted group library level features and the extracted sequence level features; training the initial prediction model by using the training set, and screening out group library level features and sequence level features which need to be reserved; respectively inputting the screened group library level features and sequence level features into the initial prediction model to obtain an optimized prediction model; and performing performance evaluation on the test set by using the optimized prediction model. According to the method, disease associated information implied in a high-diversity antibody group library can be effectively mined, and the diagnosis precision of a prediction model is effectively improved.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence platform construction, in particular to a non-invasive diagnosis method for infectious diseases based on deep learning. Background technique [0002] Early diagnosis of various infectious diseases is a key means to effectively control sudden disease pandemics. For example, the symptoms caused by 2019-nCoV infection are very similar to other respiratory diseases, making its diagnosis extremely complicated. Due to its high infectivity and fatality rate, rapid and accurate diagnostic methods are of great significance for the control of the pandemic and the treatment of patients. Other meningococcal causes, such as sepsis, have early symptoms similar to those of more self-limiting viral diseases, and delays in diagnosis and treatment can lead to death or serious complications. Therefore, screening biomarkers that can accurately distinguish the source of disease infection is t...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G16H50/80G16H50/20G16B30/00G06N3/08G06N3/04G06K9/62
CPCG16H50/80G16H50/20G06N3/08G16B30/00G06N3/045G06F18/214
Inventor 张镇海余学清陈渊蓝春红张艳芳
Owner GUANGDONG GENERAL HOSPITAL