Affordable Access

Discussion on the meeting on ‘Statistical modelling and analysis of genetic data’

Publication Date
  • Computer Science


 2002 Royal Statistical Society 1369–7412/02/64737 J. R. Statist. Soc. B (2002) 64, Part 4, pp. 737–775 Discussion on the meeting on ‘Statistical modelling and analysis of genetic data’ David J. Balding .Imperial College School of Medicine, London/ I extendmy apologies to the authors that an unavoidable commitment arising unexpectedly in the 18 hours before the meeting robbed me of my final preparation, so that my comments at the meeting were poorly presented. I shall try to do a better job in this written version, and to leave enough space I shall omit my introductory comments about the role of statisticians in bioinformatics. Broman and Speed I have no special expertise in this area, and so the authors should feel more than usually free to ignore my own views. However, I did consult widely among statistical researchers who have worked in quantitative tract locus (QTL) mapping, and I found little enthusiasm for the paper. There is not much new here for the general statistician and, as far as I could detect, there is also little of interest for the applied researcher. I have no major quibble with the paper’s analyses. I would have liked to have seen more attention paid to the multiple-QTL mapping method of Jansen (1994) rather than, or in addition to, composite interval mapping (CIM), since it attempts to tackle the ‘key problem with CIM’. Given the recent interest in Baye- sian methods for QTLmapping it would have been useful to examine them. However, my main complaint is with the paper’s emphasis, summarized by the four ‘key points that [the authors] wish to make in this paper’, listed at the start of Section 5. The authors are prominent statisticians who, I believe, do not work principally in QTL mapping. In emphasizing model selection, model comparison and simulation studies, as if they were novel ideas in the field, and drawing four pedestrian conclusions, they seem to me to con- demn the field as a statistical backwater. No doubt there is bad statistical practice in this

There are no comments yet on this publication. Be the first to share your thoughts.