Welcome to LUADEXPRESS

Expression

Users can browse the expression alterations of a given gene across different studies and identify consensus gene signatures in LUAD.

Survival

Users can explore whether the gene of interest is associated with the prognosis of LUAD patients in different datasets.

Correlation

Users can check the correlation between any two given genes.

Dependency

Users can explore whether the gene of interest in essential for the survival of LUAD cells based on genome-wide CRISPR-Cas9 loss-of-function screening data from DepMap database.

Diagnosis

Users can explore the diagnostic value of genes based on the whole blood microarray data of LUAD patients.

Prognosis

Users can build gene markers for prognosis assessment using LASSO Cox regression analysis and validate the marker on other studies.

PD-L1

Users can explore expression data of patients treated with PD-L1 inhibitor Atezolizumab.

Contact Us

Gene Expression Alteration




This feature enables users to check the expression alteration of a given gene across 11 independent LUAD datasets from GEO and TCGA databases. The horizontal dashed line represents adjusted p-value = 0.05

  • Gene: Input a gene of interest.


    • Consensus Gene




    This feature enables users to browse consensus up-regulated or down-regulated genes across 11 independent LUAD datasets from GEO and TCGA databases.

  • Direction of alteration: Select consensus up-regulated or down-regulated genes.
  • Log2 (Fold change): Set custom Log2 fold change threshold. Default value is 0
  • Adjusted p-value: Set custom adjusted p-value threshold. Default value is 0.05.



    • Survival Curve




    This feature draws overall survival (OS) or recurrent free interval (RFS) related survival curves, using Log-rank test for hypothesis test. Users can select a suitable threshold for splitting the high-expression and low-expression groups. The hazard ratio and the 95% confidence interval information based on the group cutoff will also be included in the survival plot.

  • Gene: Input a gene of interest.
  • Dataset: Select a LUAD dataset.
  • Method: Select OS or RFS for analysis.
  • Group Cutoff Method: Select a method for splitting the high-expression and low-expression groups. For maximal separation method, the cut-off point was determined based on the maximally selected log-rank statistics, meanwhile, each group will contain at least 30% of the total population to avoid assigning too few patients into a given group.
  • Cutoff-High(%): Samples with expression level higher than this cutoff are considered as the high-expression group.
  • Cutoff-Low(%): Samples with expression level lower than this cutoff are considered the low-expression group.
  • High Group Color: Set the color of high-expression group.
  • Low Group Color: Set the color of low-expression group.






    • Correlation Analysis




    This feature performs pair-wise correlation analysis to explore the correlation between two genes using methods including Pearson, Spearman and Kendall.

  • Gene A: Input a gene of interest.
  • Gene B: Input another gene of interest.
  • Dataset: Select a LUAD dataset.
  • Correlation Coefficient: The method for calculating the correlation coefficient.




    • Dependency Analysis




    This feature enables users to check the dependency of a specific gene across LUAD cell lines based on genome-wide CRISPR-Cas9 loss-of-function screening data provided by DepMap database.

  • Gene Input a gene of interest.


    • Diagnostic Gene Signature Assessment




    Molecular biomakers for early diagnosis of LUAD is urgently needed. Many studies suggested using the differentially expressed genes identified in the tumor tissues for diagnosis, but the problem is whether these genes exhibited the same level of alternations in peripheral blood? Thus, this feature enables users to explore and validate the performance of a combination of genes as the diagnostic marker for LUAD based on the valuable whole blood microarray data of 154 LUAD patients (GSE20189). This feature generates a plot displaying the Area Under the Curve (AUC) by combining all input genes using logistic regression. Also, the AUC for each gene is also displayed, along with the differentially expressed genes in this cohort.

  • Gene List: Copy and paste a list of genes into the box.
  • Color: Set the color of AUC plot.


    • LASSO Prediction Tool Construction and Assessment




    This feature enables users to construct prognosis (overall survival) prediction tool based on gene signatures using LASSO Cox regression analysis (10-fold cross validation), which is a commonly used method to reduce dimentionality and bulit prediction tool. Three datasets are available for model construction, i.e., the Meta-cohort, TCGA cohort, GSE72094 cohort. Users need to paste a list of genes (e.g., m6A genes, immune genes, metabolism genes, etc.,) into the box. if no gene is assigned with a LASSO coefficient, the lambda value may need to be adjusted or the genes enrolled may not appropriate for prognosis prediction.

  • Gene List: Copy and paste a list of genes into the box.
  • Dataset: Select a dataset for model construction. The remaining two datasets will be used as validation sets.
  • Lambda: Select a Lambda value for model construction. Lambda.min means minimum error, Lambda.1se means error is within 1 standard error of minimum error.




    • Clinical Response Gene Signature Assessment




    This feature enables users to explore the ability of using gene signatures to predict the response of anti-PD-L1 treatment. A plot displaying the Area Under the Curve (AUC) by combining all input genes using logistic regression is shown. The differentially expressed genes by comparing responders vs. non-responders in also shown.

  • Gene List: Copy and paste a list of genes into the box.
  • Color: Set the color of AUC plot.


    • Survival Curve




    This feature draws overall survival (OS) related survival curves in the anti-PD-L1 treatment cohort, using Log-rank test for hypothesis test. Users can select a suitable threshold for splitting the high-expression and low-expression groups. The hazard ratio and the 95% confidence interval information based on the group cutoff will also be included in the survival plot.

  • Gene: Input a gene of interest.
  • Group Cutoff Method: Select a method for splitting the high-expression and low-expression groups. For maximal separation method, the cut-off point was determined based on the maximally selected log-rank statistics, meanwhile, each group will contain at least 30% of the total population to avoid assigning too few patients into a given group.
  • Cutoff-High(%): Samples with expression level higher than this cutoff are considered as the high-expression group.
  • Cutoff-Low(%): Samples with expression level lower than this cutoff are considered the low-expression group.
  • High Group Color: Set the color of high-expression group.
  • Low Group Color: Set the color of lwo-expression group.



    • Quantifying immune-infiltrating cells using ssGSEA



    This section allows users to apply Single-Sample Gene Set Enrichment Analysis to quantify immune-infiltrating cells from transcriptomic data.

  • File Requirements:
  • For Expression matrix
  • Tsv/Csv/Txt format supported.
  • Rows (Gene Symbols) x Columns (Samples), duplicated gene symbols are not allowed in the rows.
  • Unit: TPM or FPKM or Microarray value.
  • For Customized Immune Cell Markers
  • Tsv/Csv/Txt format supported.
  • First Column: Immune Cell Type; Seconde Column: Gene Markers.
  • The format of the customized immune cell markers file should be identical with the example file. Please also rename the column names of your file to “CellType” and “Symbol”.

    • Download Example Data

    Click To Go

    Expression Matrix

    Results

    Download

    Batch Correction




    This section allow users to adjust batch effects prior to subsequent analysis.

  • File Requirements:
  • Expression matrix (Tsv/Csv/Txt format).
  • First column should be the sample information (e.g. Sample ID).
  • The input data are assumed to be cleaned and normalized before batch effect removal.
  • For each expression matrix, duplicated names are not allowed in the first row.
  • For each expression matrix, samples should be ordered as (control/control/control/treatment/treatment/treatment; not control/treatment/control/treatment) from left to right.

    • Download Example Data

    Please specify the number of control cases (Group 1) and the number of treatment cases (Group 2) in each expression matrix.

    Click To Go