Direct Interaction Network Inference with Compositional Data via Lasso Penalized D-Trace Loss
a) The R code in folder "\CDtr\CDTr" or "CDTr\CDtracesim\CDTr" is an implementation of CDTr in our manuscript.
b) The folder "CDTr\CDtracesim\bioinfo_CDtrace" an implementation of CDtrace, which is available at https://github.com/coamo2/CD-trace.
c) The folder "CDTr\CDtracesim\SpiecEasi-master" is an implementation of S-E(glasso) and S-E(mb) (Kurtz et al, 2015), which is available at https://github.com/zdk123/SpiecEasi by Kurtz et al (2015).
d) The folder "CDTr\CDtracesim\gcoda-master” is an implementation of gCoda (Fang et al, 2017), which is available from https://github.com/huayingfang/gCoda by Fang et al (2017).
e) The R files named "sim.R" and "simfunc" are for simulations under 12 different scenarios. The networks are estimated according to CDTr, CD-trace, gCoda, S-E(glasso) and S-E(mb) with the help of above-mentioned codes.