How do scientists use "big data" to help human disease research?

In recent years, scientists have made great progress in improving human health and helping human disease research relying on "big data" In this paper, the editor will sort out the recent relevant research results and share them with you! Photo source: cc0 public domain [1] front immune: big data helps design better influenza vaccine doi: 10.3389/fimmu.2019.02616 every year, thousands of Americans get sick because of influenza, thousands of people are hospitalized or die Influenza vaccination can reduce the chance of infection However, according to the Centers for Disease Control and prevention, the vaccine is only effective for 40% to 60% of the time at best Recently, researchers from Michigan State University have data on how RNA levels in cells change after infection or vaccination Their work can help make future influenza vaccines work better, and even help design universal vaccines Understanding these differences may help us identify new goals for building better vaccines and help us find better ways to treat the disease, the researchers said The researchers analyzed data from 18 previously published studies, in which scientists collected blood samples from influenza patients and vaccinators, and studied the gene expression of these samples The motivation of combining different data is that usually smaller data will not be statistically sufficient to detect significant differences 【2】 JACC: data of 500000 people: too much or too little sleep will increase the risk of heart attack! Doi: 10.1016/j.jacc.2019.07.022 a new study conducted by the University of Colorado Boulder on nearly 500000 people shows that even if you don't smoke and exercise regularly, there is no genetic predisposition to cardiovascular disease, and insufficient or excessive sleep will increase your risk of heart disease This study is published in Journal of the American In the journal College of Cardiology, the study also found that for those with a high genetic risk of heart disease, sleeping six to nine hours a night offsets the risk The researchers said that this provides some of the strongest evidence that sleep time is a key factor affecting heart health, which is applicable to everyone The researchers analyzed gene information, self-reported sleep habits and medical records of 461000 biobank participants aged 40-69 who have never had a heart attack, and then conducted For seven years People who slept less than six hours a night were 20 percent more likely to have a heart attack during the study period than those who slept six to nine hours a night Those who slept more than nine hours were 34 percent more likely 【3】 Nat commun: big data helps scientists to clarify the molecular mechanism of breast cancer diffusion doi: 10.1038/s41467-019-10148-6 now scientists can use a large number of genome data to detect the disease status of patients as early as possible, but it can also help researchers better study the molecular mechanism of disease development Recently, in a research report published in the international journal Nature communications, scientists from Michigan State University determined a better model to fight against the spread of breast cancer and help test potential drugs through big data analysis The current methods used in the laboratory usually use cell lines or cells cultured on plates to simulate tumors in patients' bodies The most common cause of cancer-related death is growth and diffusion, which is directly related to the death of 90% of cancer patients Up to now, few drugs can treat cancer metastasis of patients, and researchers do not know which link is wrong in the process of drug development The differences between cell lines and tumor samples may raise a number of key questions, such as the extent to which cell lines can capture tumors, said researcher Bin Chen In order to answer this question, the researchers conducted a large-scale integrated analysis on the data from multiple genomic databases, including the human cancer genome map project, cancer cell line encyclopedia, gene expression synthesis, genotype and phenotype database, etc Using these genomic data to find new cancer therapies is the ultimate goal of researchers Before researchers put a lot of costs into experiments, we need to evaluate the early research model and select the most appropriate drug detection method based on genomic characteristics 【4】 JAMA inter Med: more than 300000 people's Congress data show that following a healthy plant-based diet can reduce the risk of type 2 diabetes doi: 10.1001/jamanternmed.2019.2195 recently, an international journal published in JAMA internal In a new meta-analysis, researchers from the Chen Zengxi School of public health at Harvard University found that people who stick to a plant-based diet have a lower risk of type 2 diabetes than those who don't The researchers also found that the link was stronger for those who ate a diet that emphasized healthy plant-based foods Plant based diets have become more and more popular in recent years, so we think it is important to quantify the overall relationship between them and the risk of diabetes, especially the food ingredients with significant dietary differences, although previous studies have shown that plant-based diets may help reduce type 2 diabetes But lack of overall analysis of epidemiological evidence According to the researchers, the current study provides the most comprehensive evidence to date of the relationship between maintaining a healthy plant-based diet and reducing the risk of type 2 diabetes 【5】 Nat genetics: big data revealed that five genetic loci affecting the occurrence of Alzheimer's disease doi: 10.1038/s41588-019-0358-2 came from more than 94000 individuals Analysis of genetic data showed that there are five new risk genes for Alzheimer's disease, and 20 other known genes were identified These new findings support the development of evidence that genomes related to specific biological processes, such as cell transport, lipid transport, inflammation and immune response, are "genetic centers" and important components of disease processes, and the study is published in the journal Nature Genetics The researchers said that this ongoing collaborative study on the genetic basis of Alzheimer's disease has enabled us to delve into the complexity of this devastating disease, and as we continue to better understand and target methods to treat and prevent Alzheimer's, the scale of this study provides additional clarity for prioritized genes In the study, researchers analyzed rare and common genetic variations in 94437 patients with late-onset Alzheimer's disease, the most common form of dementia in the elderly IGAP consists of four alliances in the United States and Europe that have been collaborating on genome-wide association studies (GWAS) since 2011, involving thousands of DNA samples and shared datasets GWAS aims to detect genomic variations associated with Alzheimer's disease Understanding genetic variation is helping researchers identify the molecular mechanisms that affect the occurrence and development of disease Photo source: University of Utah Health [6] PLoS One: big data helps scientists quickly detect neonatal pyemia doi: 10.1371/journal.pone.0212665 researchers and doctors and scientists test machine learning models in the population of neonatal intensive care unit (NICU) by using data routinely collected from electronic health records (EHR) Research shows that automated procedures can quickly identify which infants in NICU have sepsis before clinical symptoms appear The results were recently published in the journal PLoS One Sepsis begins with bacteria invading the blood, and it is the leading cause of infant mortality and incidence rate worldwide Unfortunately, the aggressive immune response can lead to the development of sepsis and shock, leading to multiple organ failure, sometimes even fatal Although relatively rare in healthy infants, the incidence of sepsis is 200 times higher in preterm or chronically hospitalized infants Survivors of infant septicemia may have sequelae, such as chronic lung disease and neurodevelopmental disorders 【7】 Nat commun: scientists have successfully used big data analysis to identify the new cancer risk gene doi: 10.1038/s41467-018-04900-7 There are many genetic reasons for inducing cancer, such as some mutations will be inherited from parents, while others are acquired mutations, such as caused by external factors or DNA replication errors Large scale genome sequencing has made a series of research results in identifying cancer caused by somatic mutations, but this technology can not effectively identify cancer caused by genetic mutations And the main source of identifying these genetic mutations is still based on familial research Recently, in a research report published in the international journal Nature communications, scientists from Barcelona genomics research center have developed a new statistical method, which can identify cancer susceptibility genes from tumor sequencing data This method uses the old idea that cancer genes usually need "two strikes" (two strikes Hits) can cause cancer The researchers say this method can help them systematically identify the relevant genes from the current cancer genome database 【8】 Nat Genet: big data reveals the gene association of disease doi: 10.1038/s41588-018-0294-6 imagine the scene of rain falling on the sidewalk Although raindrops seem to fall at random, there will still be an area of the sidewalk that will remain dry over time The analogy is similar to a new approach designed by researchers at the University of Utah's Institute of health They explored more than 100000 healthy people to identify areas of our genes that are not easily altered They believe that DNA mutations in these "restricted" areas could lead to serious pediatric diseases The researchers said that we are not focusing on the place where DNA changes, but looking for the gene part where DNA remains unchanged Our model searches for exceptions to the rule of dense genetic variation in this huge data set to reveal the limited area of genes lacking variation We believe that these regions may be lethal or cause extreme phenotypes of the disease when mutated Although this approach is conceptually simple, it has only recently had enough human genome to implement it These new, unchanging extensions may reveal new pathogenic genes and can be used to help determine the cause of disease in patients with developmental disorders The relevant research results are published in the international journal Nature Genetics Previously unrelated genes often contain one or more highly restricted regions where mutations can cause disease, the researchers note 【9】 Cancer immune res: scientists are expected to use big data analysis to predict the response of cancer patients to immunotherapy doi: 10.1158/2326-6066.cir-18-0129 now in the era of big data, cancer researchers are looking for a variety of new methods to monitor the effectiveness of cancer immunotherapy; recently, an international journal published in cancer immunology In the research report, researchers from Johns Hopkins University School of medicine developed a Bioinformatics Based