From fffadbb4e3e06b245a8b56bb3d918c58e44f2a67 Mon Sep 17 00:00:00 2001 From: Matt Lane Date: Thu, 29 May 2025 20:35:57 -0700 Subject: [PATCH] Update README.md Updated README to illustrate GRN-LOOPy is updated iRF LOOPy --- README.md | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) diff --git a/README.md b/README.md index e5c8f88..ec61bfa 100644 --- a/README.md +++ b/README.md @@ -1,11 +1,18 @@ -# HP Network Generator: -This project contains code for the high performance generation of networks from machine learning methods. +# GRN-LOOPy +GRN-LOOPy is an updated version of teh iRF_LOOPpy software package. It is a high-performance Python package for inferring Gene Regulatory Networks (GRNs) using Iterative Random Forest (iRF), and now offers the options for hot swapping additional models: Random Forest, Decision and Regression Trees, and Gradient Boosted Decision Trees via the LightGBM package. This tool leverages machine learning to predict regulatory interactions between genes, generating Predictive Expression Networks (PENs). + +Features +- Efficient Network Inference: Uses iRF, an advanced tree-based model, to infer regulatory relationships from gene expression data. +- MPI-Based Task Farm: Implements a dynamic task allocation system for parallelized computation, reducing idle times compared to traditional batch queuing. +- Streamlined Workflow: Includes preprocessing, processing, and post-processing steps for ease of use. +- Consolidated Output: Unlike previous versions, all model results are stored in a single output file, improving data management. ## Installation To run this code, an system with MPI installed is required. When running this code on a slurm system, it is recommended that users refer to the system documentation and proper usage of MPI within that environment (i.e. specific lmod modules required to properly run MPI). Ex. Running the package code on OLCF's Frontier system requires + ```bash module load rocm module load boost @@ -47,12 +54,10 @@ module load cray-mpich/8.1.23 source activate /ccs/home/lanemj/environments/frontier/lightgbm export MPI4PY_RC_RECV_MPROBE=0 -srun -n 16 --gpus-per-task=1 --gpu-bind=closest python /ccs/home/lanemj/hp_network_generator/src/process.py\ - --infile /ccs/home/lanemj/DREAM5_SysGenA100_Network1_Expression_100.tsv \ +srun -n 16 --gpus-per-task=1 --gpu-bind=closest python ./src/process.py\ + --infile ../data/DREAM5_SysGenA100_Network1_Expression_100.tsv \ --device gpu \ --header_row_idx 0 \ 2> err 1> out ``` - -