Dear all, I make extensive use of the gls function from the nlme library for microarray data analysis (details described in http://bioinf.wehi.edu.au/smawehi/library/smawehi/html/gls.series.html). A typical fit involves running gls about 15000 times with the same formula, with different reponses and weights but with the same design matrix. Is there a way to avoid parsing the same formula every time gls is called by giving gls the response vector, weights and design matrix directly? That is, is there some way to call gls which is analogous to lm.fit for lm or to glm.fit for glm? Thanks for any suggestions Gordon --------------------------------------------------------------------------------------- Dr Gordon K Smyth, Senior Research Scientist, Bioinformatics, Walter and Eliza Hall Institute of Medical Research, 1G Royal Parade, Parkville, Vic 3050, Australia Tel: (03) 9345 2326, Fax (03) 9347 0852, Email: smyth at wehi.edu.au, www: http://www.statsci.org