DO NOT USE --- Codes and key processed files for Xie et al. manuscript
收藏资源简介：
## Codes and scripts used in Xie et al. manuscript #### IMPORTANT NOTE: the biomarkers and algorithms in this repo are protected by multiple patents. The codes could only be used for academic evaluations, distributed under the [CC BY-NC-ND 4.0](https://creativecommons.org/licenses/by-nc-nd/4.0/ "CC BY-NC-ND") license. Please contact Dr. Kun Sun (sunkun@szbl.ac.cn) for enquiries about commercial usages. Our programs/codes were developed on a CentOS v7.5 (64-bit, kernal version v3.10.0) machine with the following software and packages: - [Ktrim v1.6.0](https://github.com/hellosunking/Ktrim/releases/tag/v1.6.0) - [Bowtie2 v2.3.5.1](https://github.com/BenLangmead/bowtie2/releases/tag/v2.3.5.1) - [Bedtools v2.29.2](https://github.com/arq5x/bedtools2/releases/tag/v2.29.2) - [R v4.1.2](https://cran.r-project.org/bin/windows/base/old/4.1.2), requires "ggplot2", "ggpubr", "ggprism", "reshape2", "forcats", "ggbeeswarm", "dplyr", "cowplot", "tidyverse", "verification", "gbm", "caret", "foreach", "doParallel", "binom", and "pROC" packages. ### CfDNA data processing For raw cfDNA read processing, including preprocessing, alignment, and duplicate removal, please refer to [Liu et al. STAR Protocols 2025](https://www.sciencedirect.com/science/article/pii/S2666166725001637) and this [github repo](https://github.com/hellosunking/STAR-Protocol-cfDNA-end-selection). ### Feature extractions 1. For size, end selection and motif features, please refer to [Ju et al. Cell Reports Methods 2024](https://www.sciencedirect.com/science/article/pii/S2667237524002601), [Liu et al. STAR Protocols 2025](https://www.sciencedirect.com/science/article/pii/S2666166725001637), and this [github repo](https://github.com/hellosunking/STAR-Protocol-cfDNA-end-selection). 2. For E-index, please refer to [An et al. Nature Communications 2023](https://www.nature.com/articles/s41467-023-35959-6) and [this github repo](https://github.com/hellosunking/STAR-Protocol-cfDNA-end-selection). 3. Programs for copy number analysis, including extracting mean and s.d. values from the reference panel, calculating read counts per bin and converting to Z-scores, are provided in `Copy.number` directory. We first analyzed the referece panel to calculcate mean and sd values for each 5Mb genomic bin using `1.calc.mean.sd.in.reference/process.reference.sh`, and generated `1.calc.mean.sd.in.reference/Prostate.reference.bin5M.mean.sd` file; then we calculated the Z-scores for each sample per bin using `2.calc.z.score/calc.Zscore.sh`. 4. Programs for analyzing AR-specific binding regions, including region definition and extracting cfDNA signals, are provided in `AR-specific.binding.regions` directory. You need to prepare a `bed.list` file and then run `2.profile.cfDNA/profile.cfDNA.sh` to get the features. ### Building PROTECTOR and PROTECTOR+ models Programs and feature matrices for building PROTECTOR and PROTECTOR+ models are provided in `PROTECTOR.model` directory. Run `build.PROTECTOR.models.sh` to build the models and plot ROC curves.
