In response to the editorial by Halloran and colleagues {1} the CARGO study investigators wish to strongly endorse the concerns expressed regarding methodological challenges posed by genomic translational research. These include use of imperfect clinical gold standards, focused versus whole genome screening, experimental and biological variability and uncertainty about clinical replication of such findings. The large number of variables assessed in mi-croarray studies creates significant challenges that may result in false conclusions about the relationships of genes (or proteins) to disease states, as shown in the Lancet article referenced (2). However, the approach taken in the CARGO study (3) follows the outline described in a recent review (4) for addressing these issues in the development of clinically useful classifiers: the putative genes discovered with mi-croarrays (or found in the literature) were first confirmed by real-time PCR, then used to develop a classifier, which was validated using an independent sample and patient set. in this regard, the level of rigor in the CARGO study applied to algorithm development and validation in fact exceeded the recommendations of the referenced Lancet article, as described in detail in the Supplementary materials. Additional studies using different patient cohorts are ongoing to further support and extend these results, not only to the U.S. population, but also to the international adult heart transplant population.
展开▼
