Findr (Fast Inference of Networks from Directed Regulations) is a statistical inference tool for genetics. It predicts the probability of pairwise gene regulation probability based on gene expression level data. When genotype data is available for the best eQTLs, higher prediction accuracy can be achieved. The pairwise regulation probability is then applied for the reconstruction of gene regulation networks.

Findr obtains much higher accuracy and faster speed than existing programs [1]. This is due to the analytical and implementational advances. Findr's unprecedented speed allows for whole-transcriptome causal network reconstruction, with a tutorial in [2]. Findr library can be downloaded from [3].

This package is the C implementation of Findr library. It requires recent builds of GCC, GNU make, and GNU Scientific Library (GSL). Users can use the provided binary and python interfaces, or R package to interact with Findr library to perform calculations, or write one's own program and call Findr. The binary, python, and R entry points can be downloaded from [4], [5], [6] respectively. On Windows, we recommend building and running Findr on "Bash on Windows" [7], rather than building everything natively from scratch.

A more detailed documentation of Findr can be found as doc.pdf.

[1] Lingfei Wang and Tom Michoel (2017) Efficient and accurate causal inference with hidden confounders from genome-transcriptome variation data. PLOS Computational Biology 13(8): e1005703. https://doi.org/10.1371/journal.pcbi.1005703
[2] Lingfei Wang and Tom Michoel (2017) Whole-transcriptome causal network inference with genomic and transcriptomic data. bioRxiv 213371. https://doi.org/10.1101/213371
[3] https://github.com/lingfeiwang/findr
[4] https://github.com/lingfeiwang/findr-bin
[5] https://github.com/lingfeiwang/findr-python
[6] https://github.com/lingfeiwang/findr-R
[7] https://msdn.microsoft.com/commandline/wsl/about