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 binary interface of Findr library, downloadable from [4]. It requires recent builds of GCC, GNU make, and GNU Scientific Library (GSL). It also requires Findr library to be pre-installed. The python interface and R package are also provided for Findr, downloadable from [5] and [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. For the list of updates in the binary interface, see UPDATES. For updates in functionality, see README in [3]. For examples to use the binary interface, see EXAMPLES.

[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