This website was produced as an assignment for an undergratuate course at Davidson College



          Overall I thought this was a relatively well done paper. MacBeath et al. really pulled out all the stops to verify their results each step of the way. For example on the microarrarys they spotted each PDZ domain 4 times, as well as performed the microarrays in triplicate, meaning that for each PDZ/Peptide combination there were twelve total opportunites to interact. They also used microarrays and fluorescent polarization in combination, which definitely helped their results be more accurate. They refined their model a total of three times as well. All of this really helped convince me to trust their results.

          Some criticism would be that I felt some of the figures were repetitive. For example within Figure 2 (of the paper) I felt parts A and B showed almost the same thing. Really the only difference is that A helped show the strenght of the FP confirmed positives. Those figures took up a lot of space that perhaps could have been better used. Also, while was convinced that they created a way to model PDZ domain selectivity, the model they provided in the paper is quite general, and if I actually wanted to use their model they do not provide any specific details. Also, a small thing, I did not feel like the title was representative of the actual paper. I am not sure what they are trying to get at with "optimized across the mouse proteome" but many of the training set peptides they used were derived from receptor families and not even actual mouse proteins.



MacBeath, Gavin et al. PDZ Domain Binding Selectivity is Optimized Across the Mouse Proteome. 2007. Science 317: 364-369.

All figures come from the paper cited above unless otherwise noted. Figure numbers in the paper review do not match those of the paper, and some figures appear out of the original paper order. The protein structures in the PDZ header come from PDB.




Genomics            Davidson College

Email me: