iPRISM User Guide
Introduction
Welcome to the vignette for the PRISM package. This document provides an overview of the package’s functionalities, including basic usage examples and detailed explanations of the main functions. The PRISM package includes the core function for the paper named PRISM: Predicting Response to cancer Immunotherapy through Systematic Modeling.
Example 1: Correlation Plot with Cell Types and Pathways
The cor_plot function generates a correlation plot
between cell types and pathways, displaying correlation coefficients as
a heatmap and significant correlations as scatter points.
# Read cell line and pathway information
data(data.path, package = "iPRISM")
data(data.cell, package = "iPRISM")
# Draw the plot
cor_plot(data1 = data.path, data2 = data.cell, sig.name1 = "path", sig.name2 = "cell")
Example 2: Enrichment Analysis using Multiplex Networks
The get_gsea_path function constructs a multiplex
network, performs random walk restart, and calculates gene scores. It
then transforms the scores and applies GSEA using the provided gene
sets.
# Load example data
data(Seeds, package = "iPRISM")
data(ppi, package = "iPRISM")
data(path_list, package = "iPRISM")
# Shrink pathway list to the top 2 pathways
path_list <- path_list[1:5]
# Perform GSEA
result <- get_gsea_path(seed = Seeds, network = ppi, pathlist = path_list, gsea.nperm = 100)
#> Warning in asMethod(object): sparse->dense coercion: allocating vector of size
#> 1.9 GiB
print(result)
#> ES pval
#> 2-LTR circle formation 0.67796 0.06
#> 5-Phosphoribose 1-diphosphate biosynthesis -0.52301 0.31
#> A tetrasaccharide linker sequence is required for GAG synthesis 0.61996 0.00
#> ABC transporter disorders 0.55422 0.00
#> ABC transporters in lipid homeostasis -0.31617 0.15
#> padj
#> 2-LTR circle formation 0.1000
#> 5-Phosphoribose 1-diphosphate biosynthesis 0.3100
#> A tetrasaccharide linker sequence is required for GAG synthesis 0.0000
#> ABC transporter disorders 0.0000
#> ABC transporters in lipid homeostasis 0.1875Example 3: Fit Logistic Regression Model
The get_logiModel function fits a logistic regression
model as the paper highlighted, with an option for stepwise model
selection.
# Load example data
data(data_sig, package = "iPRISM")
# Fit logistic regression model
b <- get_logiModel(data.sig = data_sig, pred.value = pred_value, step = TRUE)
summary(b)
#>
#> Call:
#> glm(formula = pred ~ `Downstream signaling of activated FGFR2` +
#> `Downstream signaling of activated FGFR3` + `Downstream signaling of activated FGFR4` +
#> `FGFR1 mutant receptor activation` + `FRS-mediated FGFR3 signaling` +
#> `IGF1R signaling cascade` + `Insulin receptor signalling cascade` +
#> `MAPK3 (ERK1) activation` + `Negative regulation of FGFR1 signaling` +
#> `Negative regulation of FGFR2 signaling` + `Negative regulation of FGFR4 signaling` +
#> `PI-3K cascade:FGFR2` + `PI-3K cascade:FGFR4` + `PI3K Cascade` +
#> `Senescence-Associated Secretory Phenotype (SASP)` + `Signaling by FGFR3 in disease` +
#> `VEGFR2 mediated cell proliferation`, family = "binomial",
#> data = data.sig)
#>
#> Coefficients:
#> Estimate Std. Error z value
#> (Intercept) -29.309 10.083 -2.907
#> `Downstream signaling of activated FGFR2` 2256.687 1178.118 1.916
#> `Downstream signaling of activated FGFR3` 307.068 77.157 3.980
#> `Downstream signaling of activated FGFR4` -2258.365 1087.384 -2.077
#> `FGFR1 mutant receptor activation` 27.001 10.886 2.480
#> `FRS-mediated FGFR3 signaling` -39.058 18.304 -2.134
#> `IGF1R signaling cascade` 62.010 31.597 1.963
#> `Insulin receptor signalling cascade` 76.707 31.676 2.422
#> `MAPK3 (ERK1) activation` -21.778 8.678 -2.509
#> `Negative regulation of FGFR1 signaling` -50.247 25.410 -1.977
#> `Negative regulation of FGFR2 signaling` -1286.734 665.590 -1.933
#> `Negative regulation of FGFR4 signaling` 1251.676 613.134 2.041
#> `PI-3K cascade:FGFR2` -812.712 446.111 -1.822
#> `PI-3K cascade:FGFR4` 771.506 391.080 1.973
#> `PI3K Cascade` -117.160 37.365 -3.136
#> `Senescence-Associated Secretory Phenotype (SASP)` 17.685 11.635 1.520
#> `Signaling by FGFR3 in disease` -121.106 41.528 -2.916
#> `VEGFR2 mediated cell proliferation` -16.990 7.006 -2.425
#> Pr(>|z|)
#> (Intercept) 0.00365 **
#> `Downstream signaling of activated FGFR2` 0.05543 .
#> `Downstream signaling of activated FGFR3` 6.9e-05 ***
#> `Downstream signaling of activated FGFR4` 0.03781 *
#> `FGFR1 mutant receptor activation` 0.01312 *
#> `FRS-mediated FGFR3 signaling` 0.03286 *
#> `IGF1R signaling cascade` 0.04970 *
#> `Insulin receptor signalling cascade` 0.01545 *
#> `MAPK3 (ERK1) activation` 0.01209 *
#> `Negative regulation of FGFR1 signaling` 0.04799 *
#> `Negative regulation of FGFR2 signaling` 0.05321 .
#> `Negative regulation of FGFR4 signaling` 0.04121 *
#> `PI-3K cascade:FGFR2` 0.06849 .
#> `PI-3K cascade:FGFR4` 0.04852 *
#> `PI3K Cascade` 0.00172 **
#> `Senescence-Associated Secretory Phenotype (SASP)` 0.12851
#> `Signaling by FGFR3 in disease` 0.00354 **
#> `VEGFR2 mediated cell proliferation` 0.01531 *
#> ---
#> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#>
#> (Dispersion parameter for binomial family taken to be 1)
#>
#> Null deviance: 163.343 on 120 degrees of freedom
#> Residual deviance: 97.841 on 103 degrees of freedom
#> AIC: 133.84
#>
#> Number of Fisher Scoring iterations: 6