Reconstructing Yeast Gene Networks from Microarray Data Grace S. Shieh, Institute of Statistical Science, Academia Sinica, Taipei 115, Taiwan email: gshieh@stat.sinica.edu.tw A linear difference equations approach has been applied to reconstructing yeast gene networks from microarray data (Spellman et. al. 1998). We focus on 51 yeast genes related to DNA recombination and repair. The influences of gene j on gene i, Wij's have been obtained by a Pseudo-inverse algorithm and Ridge regression methods. The model has been checked by a "goodness of fit" quantity, which measures how close the estimates are to their input data. The goodness of fit is less than 0.01 for all three data sets. A network of genes with top-50 |Wij| values has been constructed for each data set. The consistency of the results across two data sets and three data sets has also been checked. |
Graybill Conference |