# Bayesian Processing of Microarray Images

Neil D. Lawrence, University of Sheffield
Marta Milo, University of Sheffield
Mahesan Niranjan, University of Southampton
Penny Rashbass
Stephan Soullier

in Neural Networks for Signal Processing XIII, pp 71-80

### Errata

• Page 7: on the bottom half of the page in Algorithm 1. Line number 6 (the first line in the repeat loop). This statement should be before the repeat loop starts.
Thanks to: Nilanjan Dasgupta

#### Abstract

Gene expression measurements quantify the level of mRNA produced from each gene. Two principal methods exist for producing slides for extracting these levels: photolithography and spotted arrays. One difficulty with the spotted array format is determining the size and location of the spots on the array. In this paper we present a Bayesian approach to processing images produced by these arrays that seeks posterior distributions over the size and positions of the spots. This enables us to estimate expression ratios and their variances. Exact inference for the model we specify is intractable; we develop an approximate inference technique which combines importance sampling with variational inference. Our technique has already been shown to be more consistent than both manual processing and another automated technique @Lawrence:variability03. Here we present large-scale results for twenty-four microarray slides each representing 5760 genes and show the dramatic effects of incorporating variance in our downsteam analysis. Software based on this algorithm is available for academic use.

  @InProceedings{lawrence-microarray03, title = {Bayesian Processing of Microarray Images}, author = {Neil D. Lawrence and Marta Milo and Mahesan Niranjan and Penny Rashbass and Stephan Soullier}, booktitle = {Neural Networks for Signal Processing XIII}, pages = {71}, year = {2003}, editor = {Christophe Molina and Tülay Adali and Jan Larsen and Marc Van Hulle and Scott Douglas and Jean Rouat}, month = {00}, publisher = {IEEE}, edit = {https://github.com/lawrennd//publications/edit/gh-pages/_posts/2003-01-01-lawrence-microarray03.md}, url = {http://inverseprobability.com/publications/lawrence-microarray03.html}, abstract = {Gene expression measurements quantify the level of mRNA produced from each gene. Two principal methods exist for producing slides for extracting these levels: photolithography and spotted arrays. One difficulty with the spotted array format is determining the size and location of the spots on the array. In this paper we present a Bayesian approach to processing images produced by these arrays that seeks posterior distributions over the size and positions of the spots. This enables us to estimate expression ratios and their variances. Exact inference for the model we specify is intractable; we develop an approximate inference technique which combines importance sampling with variational inference. Our technique has already been shown to be more consistent than both manual processing and another automated technique @Lawrence:variability03. Here we present large-scale results for twenty-four microarray slides each representing 5760 genes and show the dramatic effects of incorporating variance in our downsteam analysis. Software based on this algorithm is available for academic use.}, crossref = {Molina:nnsp03}, key = {Lawrence:microarray03}, linkpdf = {ftp://ftp.dcs.shef.ac.uk/home/neil/visMicroarray.pdf}, linkpsgz = {ftp://ftp.dcs.shef.ac.uk/home/neil/visMicroarray.ps.gz}, linksoftware = {http://inverseprobability.com/vis/}, group = {shefml,mig} }
 %T Bayesian Processing of Microarray Images %A Neil D. Lawrence and Marta Milo and Mahesan Niranjan and Penny Rashbass and Stephan Soullier %B %C Neural Networks for Signal Processing XIII %D %E Christophe Molina and Tülay Adali and Jan Larsen and Marc Van Hulle and Scott Douglas and Jean Rouat %F lawrence-microarray03 %I IEEE %P 71--80 %R %U http://inverseprobability.com/publications/lawrence-microarray03.html %X Gene expression measurements quantify the level of mRNA produced from each gene. Two principal methods exist for producing slides for extracting these levels: photolithography and spotted arrays. One difficulty with the spotted array format is determining the size and location of the spots on the array. In this paper we present a Bayesian approach to processing images produced by these arrays that seeks posterior distributions over the size and positions of the spots. This enables us to estimate expression ratios and their variances. Exact inference for the model we specify is intractable; we develop an approximate inference technique which combines importance sampling with variational inference. Our technique has already been shown to be more consistent than both manual processing and another automated technique @Lawrence:variability03. Here we present large-scale results for twenty-four microarray slides each representing 5760 genes and show the dramatic effects of incorporating variance in our downsteam analysis. Software based on this algorithm is available for academic use. 
 TY - CPAPER TI - Bayesian Processing of Microarray Images AU - Neil D. Lawrence AU - Marta Milo AU - Mahesan Niranjan AU - Penny Rashbass AU - Stephan Soullier BT - Neural Networks for Signal Processing XIII PY - 2003/01/01 DA - 2003/01/01 ED - Christophe Molina ED - Tülay Adali ED - Jan Larsen ED - Marc Van Hulle ED - Scott Douglas ED - Jean Rouat ID - lawrence-microarray03 PB - IEEE SP - 71 EP - 80 L1 - ftp://ftp.dcs.shef.ac.uk/home/neil/visMicroarray.pdf UR - http://inverseprobability.com/publications/lawrence-microarray03.html AB - Gene expression measurements quantify the level of mRNA produced from each gene. Two principal methods exist for producing slides for extracting these levels: photolithography and spotted arrays. One difficulty with the spotted array format is determining the size and location of the spots on the array. In this paper we present a Bayesian approach to processing images produced by these arrays that seeks posterior distributions over the size and positions of the spots. This enables us to estimate expression ratios and their variances. Exact inference for the model we specify is intractable; we develop an approximate inference technique which combines importance sampling with variational inference. Our technique has already been shown to be more consistent than both manual processing and another automated technique @Lawrence:variability03. Here we present large-scale results for twenty-four microarray slides each representing 5760 genes and show the dramatic effects of incorporating variance in our downsteam analysis. Software based on this algorithm is available for academic use. ER - 
 Lawrence, N.D., Milo, M., Niranjan, M., Rashbass, P. & Soullier, S.. (2003). Bayesian Processing of Microarray Images. Neural Networks for Signal Processing XIII :71-80