| This Article | |
 | |
 | Abstract |
| Full Text |
| |
| Services | |
| |
| Email this article to a colleague |
| Alert me to new issues of the journal |
| Request Copyright Permission |
| |
| Citing Articles | |
| |
| Citing Articles via CrossRef |
| ||||||||||||||||||||||||||||
|
Caspi et al. 10.1073/pnas.0704292104. |
Fig. 3. Flow chart summarizing candidate gene and marker selection strategy.
Fig. 4. The importance of FADS2 in n-6 and n-3 fatty acid pathways.
Fig. 5. FADS2 Linkage Disequilibrium (LD) map. Physical location and linkage disequilibrium pattern of SNPs covering the FADS2 gene from the promoter region to the 3' end of the gene (and also including FADS1). The top part of the panel shows the head-to-head orientation of FADS2 and FADS1, and the position of these SNPs. The bottom part of the panel shows an LD plot generated with Haploview, using r2 and D' as the measure of LD.
SI Text
Candidate Gene and Marker Selection Strategy. The flow chart (SI Fig. 3) shows the logic of our inquiry in selecting a candidate gene that may moderate the effect of breast feeding on the IQ. It summarizes the process by which we arrived at FADS2 as our candidate gene and how we selected, a priori, genetic markers for our test of a gene ´ environment interaction in predicting the IQ.
As we have discussed elsewhere (1), the soundest logical basis for selecting a candidate gene for inclusion in the study of a gene-environment interaction is evidence that the gene is related to the organism's reactivity to the environmental exposure. As such, in the first step of developing our hypothesis, we focused our attention on the putative biological process by which breast feeding is hypothesized to affect IQ. Specifically, breastfeeding is thought to influence brain development via nutritional processes involving fatty acids (2-4).
In the second step of developing our hypothesis, we sought to identify candidate genes that are involved in this aforementioned biological pathway. Specifically, we were interested in genes that were involved in the metabolism of fatty acids or in genes whose expression was affected by fatty acids. A review of the literature and the KEGG database (5) led us to FADS2 as the most obvious candidate gene for two reasons. (i) FADS2 encodes the delta6 desaturase. As shown in SI Fig. 4, delta-6 desaturation is a necessary and unavoidable step in the endogenous synthesis of LC-PUFAs from their dietary precursors. Of all of the gene products involved in this pathway, delta-6 desaturase catalyses the rate limiting step, making it the most crucial element in controlling the availability of endogenously produced AA and DHA. (ii) Expression of delta-6 desaturase is also regulated via end-product inhibition and dietary LC-PUFAs such as are available in breast milk. FADS2 was the only gene we tested.
In the third step of developing our hypothesis, we needed to select markers in FADS2. To the best of our knowledge, FADS2 polymorphisms have not been studied in relation to breast milk or to IQ. In this event, a standard approach to marker selection would be to cover the entire gene with Tag SNPs. The difficulty with this approach is that it yields many SNPs with small minor allele frequencies (MAF) that are not useable in regression analyses of gene-environment interaction as they would lead to statistical tests that would be woefully underpowered (i.e., the statistical power to detect interactions is determined, in tandem, by allele frequency and environmental exposure frequency). For this reason, we used HapMap data (version 20) to strategically select two Tag SNPs for two reasons (as shown in the LD map in SI Fig. 5 these SNPs, in the red box, are rs1535 and rs174575). (i) Compared to all of the Tag SNPs, these two SNPs had minor allele frequencies (.37 and .28, respectively) which ensured that we would have sufficient statistical power to test a hypothesis of a gene x environment interaction, and (ii) these two SNPs are located to cover variation in substantial parts (including most significantly in the promoter region) of FADS2 and also of FADS1, a gene which lies in a head-to-head orientation with FADS2 and which is also involved in fatty acid metabolism, encoding the delta5 desaturase (see SI Fig. 4). rs1535 and rs174575 were the only markers we tested.
To summarize, our hypothesis-driven research about gene x environment interaction proceeded in the following fashion: (1) Start with the environmental exposure; (2) specify the putative biological pathway by which it affects the phenotype; (3) identify a candidate gene involved in this same pathway; (4) select the most promising markers; (5) test and replicate.
1. Moffitt TE, Caspi A, Rutter M (2005) Arch Gen Psychiatry 62:473-481.
2. Institute of Medicine (2004) Infant Formula: Evaluating the Safety of New Ingredients (National Academies Press, Washington, DC).
3. Heird WC, Lapillonne A (2005) Annu Rev Nutr 25:549-571.
4. Marszalek JR, Lodish HF (2005) Annu Rev Cell Dev Biol 21:633-657.
5. Kanehisa M, Goto S, Hattori M, Aoki-Kinoshita KF, Itoh M, Kawashima S, Katayama T, Araki M, Hirakawa M (2006) Nucleic Acids Res 34:D354-D357.
| ||||||||||||||||||||||||||||