February 1st, 2017

This repository includes DeBias' Matlab source code and test data

• DeBias decouples global bias from direct interactions in co-orientation and co-localization data.
• Webserver: https://debias.biohpc.swmed.edu/
• Contact: Assaf Zaritsky - assafzar@gmail.com.

Apply DeBias:

---

1. Calculate GI and LI with DeBias for a set of observations x, y, output plots to outDname. The only required parameters are x,y,isCoOrientation, see documentation in the code regarding default values

   • [GI,LI] = executeDebias(x,y,isCoOrientation,outDname,name,k,nSimulations)

2. Execute DeBias on all files in a directory. Supports input file formats .xls, .txt (two columns with the matched x,y observations) and .mat (variables x,y with the matched observations).

   • executeDebiasDirectory(dname,isCoOrientation,k)

Simulations:

---

• Generate X,Y values through simulations and use executeDebias
  1. Co-orientation simulation (see Figure 1B):
     • [X,Y] = generateCoOrientSimulationData(muX,sX,muY,sY,alpha,N);
  2. Colocalization simulation (see Appendix 3 - figure 1A):
     • [X,Y] = generateColocSimulationData(sX,sZ,N,fracColoc)

Example data:

---

• Avaialble at the testData folder. Three experimental instances for co-orientation, one for co-localization and a simulated orientation dataset that include 4 instances from 2 different scenarios (Fig. 1C; data11-14: 𝜎X = 𝜎Y = 21, α = 0.2; data21-24: 𝜎X = 𝜎Y = 29, α = 0.4).

• The most recommended way to start working with the source code is to generate simulated data and try DeBias on it. For example:

  • [X,Y] = generateCoOrientSimulationData(0,30,0,30,0.3,1000);
  • [GI,LI] = executeDebias(X,Y,true);

DeBias Analyst:

---

• Compares multiple GI,LI values derived from two different experimental conditions.
• Outputs (GI,LI) plots (such as in Fig. 3D in the manuscript) and statistical significance of the GI and LI values across these conditions.
• See documentation in the code for details
  • [GIpval,LIpval] = DeBiasAnalyst(GIs1,LIs1,GIs2,LIs2,str1,str2,outdir);

Automatic estimation of K:

---

• Estimate K (number of histogram bins for EMD calculations) using the Freedman Diaconis rule.
• DeBias' ability to discriminate between different experiemtnal conditions are not sensitive to the choice of K
• (In the manuscript: Figure 3 - figure supplement 1D-E, Figure 6 - figure supplement 1I-J, Appendix 2 - figure 2, Appendix 3 - figure 1F).
• This function is provided for the curious user ;-)
• data - is the alignment data
  • K = DeBiasGetK(data,isCoOrientation)

---

Citation

• A webserver implementing this code is available at https://debias.biohpc.swmed.edu/, a comprehensive user manual is available at the web-site.

  • Source code of the database and for setting the webserver are also available upon request.

• Please cite the following manuscript when using the DeBias software:

  • Decoupling global biases and local interactions between cell biological variables, eLife, 2017, 6:e22323, written by Assaf ZaritskyUri ObolskiZhuo GanCarlos R ReisZuzana KadlecovaYi DuSandra L Schmid, Gaudenz Danuser.

---

Please contact Assaf Zaritsky, assafzar@gmail.com, for any questions / suggestions.

---

For more work from the Danuser Lab

(and Software)