Getting Started

Here we guide you to get started on using FISHNET for your next project.

Citation

If you use FISHNET in published research, please cite the following paper:

Acharya et al., FISHNET: A Network-based Tool for Analyzing Gene-level P-values to Identify Significant Genes Missed by Standard Methods, doi:10.1101/2025.01.29.635546.

Table of contents

1. Installation
2. Quick start
   1. Execution
   2. Output files

---

Installation

Clone the FISHNET GitHub repository and cd into the local repository:

$ git clone https://github.com/BrentLab/fishnet.git
$ cd fishnet/

Run FISHNET on test data to verify successful installation:

$ ./fishnet.sh --test

If your platform does not support Docker (e.g. in a HPC environment), use the --singularity flag:

$ ./fishnet.sh --test --singularity

Quick start

Execution

Here we show the most basic steps for a FISHNET analysis. There are a variety of configuration options which are detailed in Configuration. This code chunk assumes that you have an input summary statistics file (comma-separated with columns named “Genes” and “p_vals”) and a set of tab-delimited network module files.

example summary statistics file (data/pvals/exampleOR/maleWC/0-maleWC.csv)

Genes,p_vals
BRD3OS,0.12219547800000001
PTCH2,0.8389969
ASNS,0.274650263
TTLL9,0.111914007
AKR1D1,0.784900795
LINC00839,0.901963325

example module file format (data/modules/ker_based/ppi2.txt)

0       1.0     SPEF2   PDE4B   SERTAD3
1       1.0     CCL3L3  CCL15   CCL4L1  CCL3    CCL16   CCL5    CCL7    CCL28   CCL4L2  CCL14   CCL27   CXCL9   CCL8    CCL23   CXCL13  CCL26   CCL13   CCL24   CCL11   CCL3L1  CCL2
2       1.0     ZHX1    COPS6   PLGRKT  SCO2    ABCB8
3       1.0     PLXNC1  PHACTR2 HCN3    NEK8    SEMA7A  GPR63   GPR45   GPX1    TULP1   NCKAP5  BRPF3   FYN     GCFC2   LTBP2   LDAF1   RHOU    PAX7    SCARF2
4       1.0     C7      C8B     NKX1-2  C8G     AKR1C1  C6      MMEL1   C8A     RFFL    C5      ZBTB1
5       1.0     PIK3AP1 SLC23A1 ASAP1   CNTNAP1 SNX17   TGOLN2  CD2AP   ASAP2   ACAP1   F2RL2

To execute FISHNET, specify the --study argument with a directory path containing all trait subdirectories with summary statistics files and the --modules argument with the directory containing input network module files.

$ ./fishnet.sh \
    --study /path/to/input/study/directory/  \
    --modules /path/to/modules/directory/

This will run both stages of FISHNET with 200 permutations using the default thresholding method (by default).

Output files

The output files are saved in the results/ directory (by default) relative to the current working directory where you launch fishnet.sh. Here, we guide you through the expected output directory structure.

1. The results directory will contain a <study> subdirectory containing output files for the input study traits and a <study>RR subdirectory containing similar output files for the permutations.

   results/
   ├── <study>/
   └── <study>RR/
   

2. Each results/<study>/ directory, will look like so:

    results/
    └── <study>/
        ├── GO_summaries/
        ├── masterSummaries/
        ├── master_summary.csv
        ├── master_summary_<study>.csv
        ├── master_summary_filtered.csv
        ├── master_summary_filtered_parsed.csv
        ├── pascalInput/
        ├── pascalOutput/
        ├── results/
        └── summary/
   

   The GO_summaries/ directory lists statistics for the GO overrepresentation analysis. The FISHNET genes for the trait and the network the genes originate from can be found in summary/<network>_<trait>_fishnet_genes_<FDR-threshold>.csv

Read through the Example to interpret the output files.