CoGPS

A novel method for deciphering context-specific gene programs using single-cell and spatial transcriptomics data

CoGPS is a computational framework for characterizing context-specific gene programs within scRNA-seq and ST datasets. CoGPS leverages directed graphs and functional gene lists to construct functional networks underlying distinct cellular and spatial contexts, which further identifies context-dependent hub genes based on the topology of networks and assigns activity levels of gene programs to individual cells or spatial locations.

Installation

You can install the released version of CoGPS with:

if (!requireNamespace("devtools", quietly = TRUE))
    install.packages("devtools")
devtools::install_github("Lin-bioinfo/CoGPS")

Getting started

1. Static cellular contexts (the test data can be downloaded from this drive_link.

library(CoGPS)
library(parallel)
cores = detectCores(logical = FALSE) - 1

load("CoGPS_data_sc_discrete.RData")
dat = pre_sc(x = x, label = label, modes = c("discrete"), genelist = genelist, transition_point = NULL)

cl = makeCluster(cores)
net_A = net_sc(dat$A, modes = c("discrete"), genelist = genelist, select = 1, cluster = cl, seed = 1)
cl = makeCluster(cores)
net_B = net_sc(dat$B, modes = c("discrete"), genelist = genelist, select = 1, cluster = cl, seed = 1)

CoGPS_score_A = ident_sc_score(dat$A, net_A, seed = 1)
CoGPS_score_B = ident_sc_score(dat$B, net_B, seed = 1)
CoGPS_state_A = ident_sc_state(CoGPS_score_A)
CoGPS_state_B = ident_sc_state(CoGPS_score_B)

2. Dynamic cellular contexts

library(CoGPS)
library(parallel)
cores = detectCores(logical = FALSE) - 1

load("CoGPS_data_sc_continuous.RData")
dat = pre_sc(x = x, label = label, modes = c("continuous"), genelist = genelist, transition_point = 20)

cl = makeCluster(cores)
net = net_sc(dat, modes = c("continuous"), genelist = genelist, select = 1, cluster = cl, seed = 1)

CoGPS_score = ident_sc_score(x, net, seed = 1)
CoGPS_state = ident_sc_state(CoGPS_score)

3. Spatial contexts

load("CoGPS_data_st.RData")
st_anno = pre_st(x_counts, coords, genelist, ref, cell_state, core_number = 2)
roiID = st_anno[st_anno$CoGPS_state == "High", "spot_ID"]
x_st = net_st_pre(x_data, coords, genelist, roiID, r = 60)

cl = makeCluster(cores)
net_st = net_st(x_st$dat, genelist, select = 1, cluster = cl, seed = 1)
involvedID = c(roiID, x_st$neighborID)
CoGPS_score = ident_st_score(x_impute, involvedID, net_st)   # use data after imputation

License

CoGPS is licensed under the GNU General Public License v3.0.

Citation

Please cite the following article if you use CoGPS in your research:

Contact

Should you have any questions, please feel free to write issues for this repository or contact the author directly at [email protected].