Data

• Schizosaccharomyces pombe

The orginal data sets with missing values, described in the article of Rustici et al.⁽¹⁾, can be downloaded from the website of the Sanger Institute.

• Test data sets

The test sets that have been used for the evaluation of the DTWimpute algorithm performance can be downloaded from here Test-Data.zip.

Software

The executable of the two-pass DTWimpute algorithm can be downloaded from here DTWimpute.exe. The software provides for a choice between three different initial rough imputation methods to be applied at the first pass of the algorithm: 1) the mean expression over the respective row or 2) the average of the two non-missing row neighbours of each missing value or even 3) KNNimpute.⁽²⁾.

• Usage options:

DTWimpute <matrix> <R> <InitialImpute> <outfile>

<matrix> = matrix with the missing values
<R> = relative radius (use R between 0.1 and 10)
<InitialImpute> = initial imputation approach to use

0 = row mean; 1 = neighbours' average; 2 = KNNimpute

<outFile> = file to receive the full matrix

(1) Rustici,G., Mata, J., Kivinen, K., Lio, P., Penkett, C. J., Burns, G., Hayles, J., Brazma, A., Nurse, P., Bähler, J. (2004) Periodic gene expression program of the fission yeast cell cycle, Nature Genetics, 36, 809-817.

(2) Troyanskaya,O., Cantor,M., Sherlock,G., Brown,P., Hastie,T., Tibshirani,R., Botstein,D., Altman,R.B. (2001) Missing value estimation methods for DNA microarrays, Bioinformatics, 17, no. 6 520-525.