update

README.md

@@ -16,23 +16,38 @@ address it with scRank using a target-perturbed gene regulatory network (tpGRN)
 #devtools::install_github("rikenbit/rTensor")
 devtools::install_github("ZJUFanLab/scRank")
 ```
-
+## Note
+
+### Key Updates
+- **Disease Relevance and Drug Effects Analysis:** Introducing the new `scRank_GSEA()` and `plot_drug_function()` functions for analyzing disease relevance and drug effects in the highest-ranking cell types.
+- **Drug Type Specification:** Added a `type` parameter in `rank_celltype()` to specify modeling effects of either agonists or antagonists, enhancing the versatility of drug response modeling.
+- **Efficient Large Matrix Manipulations:** Integration of the Python module "tensorly" in the `Constr_net()` function, with a new parameter `use_py`, to optimize large-scale data processing.
+- **Enhanced Cell State Discernment:** Integration of the `scSHC` algorithm into the `CreateScRank()` function with an `if_cluster` parameter, improving the tool's ability to discern various cell states. [More about scSHC](https://github.com/igrabski/sc-SHC).
+- **Incorporating Drug Resistance Mechanisms:** The `resistance_target` parameter in `rank_celltype()` allows for inputting targets of alternative pathways, aiding in the consideration of drug resistance mechanisms.
+- **Flexible Edge Weight Adjustment:** Introduction of the `keep_ratio` parameter to adjust edge weights in the gene regulatory network, allowing for differential treatment of node types.
+
+### To-Do
+- **Packaging and Accessibility:** We are in the process of submitting scRank to Bioconductor or CRAN for enhanced accessibility.
 ## Overview
-scRank method consists of two components, wherein the first is to reconstruct the gene regulatory network from expression ptrofiles using `Constr_net` function and the second step is to estimate the effect of the in silico drug perturbation for GRNs in each cell type using `rank_celltype` function. 
+scRank method consists of two components, wherein the first is to reconstruct the gene regulatory network from expression ptrofiles using `Constr_net` function and the second step is to estimate the extent of the in silico drug perturbation for GRNs in each cell type using `rank_celltype` function. 
 
 scRank start with create a S4 object by `CreateScRank` function:
 - the `input` is the gene expression profil eand `meta` is the cell type information. 
 - `cell_type` is the column name of the cell type information in `meta` 
 - `species` is the species of the data. ("mouse" or "human")
-- `drug` is the drug name and `target` is the target gene of the drug. `drug` could be any inhibitor in our database `utile_database`. if you know the specific target gene of the drug, you can input the target gene into `target` without inputing `drug`.
+- `drug` is the drug name and `target` is the target gene of the drug. `drug` could be found in our database `utile_database`. if you know the specific target gene of the drug, you can input the target gene into `target` without inputing `drug`.
+- `type` characters meaning the MOAs of drug including antagonist or agonist. Default is antagonist.
+- `if_cluster` A logical meaning whether clustering single-cell transcriptomic data. Default is `FALSE`.
 
 ```{r}
 CreateScRank <- function(input,
                          meta,
                          cell_type,
                          species,
                          drug,
-                         target)
+                         target,
+                         type,
+                         if_cluster)
 ```
 
 The format of the `input` is as follows:
@@ -54,6 +69,7 @@ we load the demo dataset from Seurat object, the drug target is known as Brd4.
 
 ```{r}
 seuratObj <- system.file("extdata", "AML_object,rda", package="scRank")
+load(seuratObj)
 obj <- CreateScRank(input = seuratObj,
                     species = 'mouse',
                     cell_type = 'label',
@@ -73,9 +89,10 @@ obj <- scRank::rank_celltype(obj)
 the final infered rank of cell types that determine the drug response is stored in `obj@cell_type_rank`
 
 ### 4. Visualize the result
-For visulizing the rank of cell types in dimension reduction space, we can use the `plot_dim` function.
+For visulizing the rank of cell types in dimension reduction space, we can use the `plot_dim` function after `init_mod()`.
 
 ```{r}
+obj <- init_obj(obj)
 plot_dim(obj)
 ```
 <img src='https://github.com/ZJUFanLab/scRank/blob/main/img/scRank_data.png'>