Package 'PRSPGx'

Construct disease PRS unadjusted or using clumping and thresholding

Description

Shrink prognostic effect sizes by p-value cutoff (PRS-Dis-CT turns out to be PRS-Dis-Unadj when setting p-value cutoff = 1)

Usage

PRS_Dis_CT(
  DIS_GWAS,
  G_reference,
  pcutoff = 1e-05,
  clumping = TRUE,
  p1 = 1e-04,
  d1 = 250000,
  r1 = 0.8
)

Arguments

DIS_GWAS	a numeric matrix containing disease GWAS summary statistics, including SNP ID, position, $\beta$, SE($\beta$), p-value, N, and MAF
G_reference	a numeric matrix containing the individual-level genotype information from the reference panel (e.g., 1KG)
pcutoff	a numeric value indicating the p-value cutoff
clumping	a logical flag indicating should clumping be performed
p1	a numeric value indicating p-value threshold to decide flag SNPs in clumping
d1	a numeric value indicating window size in clumping
r1	a numeric value indicating correlation in clumping

Details

PRS-Dis-CT automatically sets predictive effect sizes equivalent to the prognostic effect sizes; and only need disease GWAS summary statistics

Value

A numeric list, the first sublist contains estimated prognostic effect sizes, the second sublist contains estimated predictive effect sizes

Author(s)

Song Zhai

References

Euesden, J., Lewis, C.M. & O'Reilly, P.F. PRSice: Polygenic Risk Score software. Bioinformatics 564, 1466-1468 (2015).

Zhai, S., Zhang, H., Mehrotra, D.V. & Shen, J. Paradigm Shift from Disease PRS to PGx PRS for Drug Response Prediction using PRS-PGx Methods (submitted).

Examples

data(PRSPGx.example); attach(PRSPGx.example)
coef_est <- PRS_Dis_CT(DIS_GWAS, G_reference, pcutoff = 0.01, clumping = TRUE)
summary(coef_est$coef.G)
summary(coef_est$coef.TG)

---

Construct disease PRS using LDpred2

Description

Using snp_ldpred2_grid function from bigsnpr function

Usage

PRS_Dis_LDpred2(DIS_GWAS, G_reference, pcausal, h2)

Arguments

DIS_GWAS	a numeric matrix containing disease GWAS summary statistics, including SNP ID, position, $\beta$, SE($\beta$), p-value, N, and MAF
G_reference	a numeric matrix containing the individual-level genotype information from the reference panel (e.g., 1KG)
pcausal	a numeric value indicating the hyper-parameter as the proportion of causal variants
h2	a numeric value indicating the estimated heritability

Details

PRS-Dis-LDpred2 automatically sets predictive effect sizes equivalent to the prognostic effect sizes; and only need disease GWAS summary statistics and external reference genotype

Value

A numeric list, the first sublist contains estimated prognostic effect sizes, the second sublist contains estimated predictive effect sizes

Author(s)

Song Zhai

References

Prive, F., Arbel, J. & Vilhjalmsson, B.J. LDpred2: better, faster, stronger. Bioinformatics 36, 5424-5431 (2020).

Zhai, S., Zhang, H., Mehrotra, D.V. & Shen, J. Paradigm Shift from Disease PRS to PGx PRS for Drug Response Prediction using PRS-PGx Methods (submitted).

Examples

data(PRSPGx.example); attach(PRSPGx.example)
coef_est <- PRS_Dis_LDpred2(DIS_GWAS, G_reference, pcausal = 0.1, h2 = 0.4)
summary(coef_est$coef.G)
summary(coef_est$coef.TG)

---

Construct PGx PRS using Bayesian regression

Description

Flexibly shrink prognostic and predictive effect sizes simutaneously with glocal-local shrinkage parameters

Usage

PRS_PGx_Bayes(
  PGx_GWAS,
  G_reference,
  n.itr = 1000,
  n.burnin = 500,
  n.gap = 10,
  paras,
  standardize = TRUE
)

Arguments

PGx_GWAS	a numeric list containing PGx GWAS summary statistics (with SNP ID, position, $\beta$, $\alpha$, 2-df p-value, MAF and N), SD(Y), and mean(T)
G_reference	a numeric matrix containing the individual-level genotype information from the reference panel (e.g., 1KG)
n.itr	a numeric value indicating the total number of MCMC iteration
n.burnin	a numeric value indicating the number of burn in
n.gap	a numeric value indicating the MCMC gap
paras	a numeric vector containg hyper-parameters ($v$, $\phi$)
standardize	a logical flag indicating should phenotype and genotype be standardized

Details

PRS-PGx-Bayes only needs PGx summary statistics and external reference genotype

Value

A numeric list, the first sublist contains estimated prognostic effect sizes, the second sublist contains estimated predictive effect sizes

Author(s)

Song Zhai

References

Ge, T., Chen, CY., Ni, Y. et al. Polygenic prediction via Bayesian regression and continuous shrinkage priors. Nat. Commun. 10, 1776 (2019).

Zhai, S., Zhang, H., Mehrotra, D.V. & Shen, J. Paradigm Shift from Disease PRS to PGx PRS for Drug Response Prediction using PRS-PGx Methods (submitted).

Examples

data(PRSPGx.example); attach(PRSPGx.example)
paras = c(3, 5)
coef_est <- PRS_PGx_Bayes(PGx_GWAS, G_reference, paras = paras, n.itr = 10, n.burnin = 5, n.gap = 1)
summary(coef_est$coef.G)
summary(coef_est$coef.TG)

---

Construct PGx PRS unadjusted or using clumping and thresholding

Description

Shrink prognostic and predictive effect sizes simutaneously by 2-df (main and interaction) p-value cutoff (PRS-PGx-CT turns out to be PRS-PGx-Unadj when setting p-value cutoff = 1)

Usage

PRS_PGx_CT(
  PGx_GWAS,
  G_reference,
  pcutoff = 1e-04,
  clumping = TRUE,
  p1 = 1e-04,
  d1 = 250000,
  r1 = 0.8
)

Arguments

PGx_GWAS	a numeric matrix containing PGx GWAS summary statistics, including SNP ID, MAF, position, $\beta$, $\alpha$, 2-df p-value, and N
G_reference	a numeric matrix containing the individual-level genotype information from the reference panel (e.g., 1KG)
pcutoff	a numeric value indicating the p-value cutoff
clumping	a logical flag indicating should clumping be performed
p1	a numeric value indicating p-value threshold to decide flag SNPs in clumping
d1	a numeric value indicating window size in clumping
r1	a numeric value indicating correlation in clumping

Details

PRS-PGx-CT only needs PGx summary statistics

Value

A numeric list, the first sublist contains estimated prognostic effect sizes, the second sublist contains estimated predictive effect sizes, the third sublist contains 2-df p-values

Author(s)

Song Zhai

References

Zhai, S., Zhang, H., Mehrotra, D.V. & Shen, J. Paradigm Shift from Disease PRS to PGx PRS for Drug Response Prediction using PRS-PGx Methods (submitted).

Examples

data(PRSPGx.example); attach(PRSPGx.example)
coef_est <- PRS_PGx_CT(PGx_GWAS, G_reference, pcutoff = 0.01, clumping = TRUE)
summary(coef_est$coef.G)
summary(coef_est$coef.TG)

---

Construct PGx PRS using penalized regression

Description

Shrink prognostic and predictive effect sizes simultaneously via the penalized term. With different assumptions on the relationship between the two effects, can be PRS-PGx-L (Lasso), PRS-PGx-GL (Group Lasso), and PRS-PGx-SGL (Sparse Group Lasso)

Usage

PRS_PGx_Lasso(Y, Tr, G, intercept = TRUE, lambda, method, alpha = 0.5)

Arguments

Y	a numeric vector containing the quantitative trait
Tr	a numeric vector containing the treatment assignment
G	a numeric matrix containing genotype information
intercept	a logical flag indicating should intercept be fitted (default=TRUE) or set to be FALSE
lambda	a numeric value indicating the penalty
method	a logical flag for different penalized regression methods: 1 = PRS-PGx-L, 2 = PRS-PGx-GL, 3 = PRS-PGx-SGL
alpha	a numeric value indicating the mixing parameter (only used when method = 3). alpha = 1 is the lasso penalty. alpha = 0 is the group lasso penalty

Details

PRS-PGx-Lasso requires individudal-level data

Value

A numeric list, the first sublist contains estimated prognostic effect sizes, the second sublist contains estimated predictive effect sizes

Author(s)

Song Zhai

References

Yang, Y. & Zou, H. A fast unified algorithm for solving group-lasso penalize learning problems. Statistics and Computing 25, 1129-1141 (2015).

Simon, N., Friedman, J., Hastie, T. & Tibshirani, R. Fit a GLM (or cox model) with a combination of lasso and group lasso regularization. R package version, 1 (2015).

Zhai, S., Zhang, H., Mehrotra, D.V. & Shen, J. Paradigm Shift from Disease PRS to PGx PRS for Drug Response Prediction using PRS-PGx Methods (submitted).

Examples

data(PRSPGx.example); attach(PRSPGx.example)
coef_est <- PRS_PGx_Lasso(Y, Tr, G, lambda = 1, method = 1)
summary(coef_est$coef.G)
summary(coef_est$coef.TG)

---

Simulated example data

Description

Simulated example data required by PRS-DIS and PRS-PGx functions.

Usage

data(PRSPGx.example)

Format

A list with 8 sublists:

PGx_GWAS

PGx GWAS including SNP ID, MAF, position, $\beta$, $\alpha$, 2-df p-value, and N; SD(Y), and mean(T)

DIS_GWAS

disease GWAS including SNP ID, MAF, position, $\beta$, SE($\beta$), p-value, and N

G_reference

simulated individual-level genotype from the reference panel matched with the simulated sample PGx genotype

Y

simulated phenotype (continuous)

T

simulated treatment assignment, 1 = treatment, 0 = placebo

G

simulated sample PGx genotype with 100 SNPs and 4000 subjects

beta

simulated prognostic effect sizes (i.e., the underlying true prognostic effect sizes)

alpha

simulated predictive effect sizes (i.e., the underlying true predictive effect sizes)

---