Comment re: A Comparison of DNA Copy Number Profiling Platforms

2008 
The letter from Coe and colleagues acknowledges the values ofour systematic approach in comparing DNA copy number profilingplatforms but disagrees with how resolution and coverage weredefined in the study.We agree with Coe and colleagues that no individual aspect ofplatform performance, such as resolution, can be measured as asingle metric. The actual minimal alteration size that can bereliably detected by a platform depends upon multiple variablesincluding local probe spacing, sequence complexity, and genomeploidy. In our case, the calculated effective resolution is a pointestimate of the size of DNA alteration one could expect to detectunder comparable conditions. The value of this calculation thusresides in the relationship between array density when controllingfor array platform. In this case, we found a f4.5-fold increase ineffective resolution with a similar fold increase in probe densityfor both the Affymetrix 100K/500K and Agilent 44K/185K plat-forms. The difference between our empirical estimate of 17.7 kbas the effective resolution of the Affymetrix 500K platform andthe theoretical calculation reported by Coe and colleagues (1) maybe due to fact that these studies emphasized 2-fold versus1.5-fold alterations, respectively. We also agree with Coe andcolleagues in that statistical or heuristic approaches to definingbreakpoints are advantageous. As such, we used a segmentation-based approach (2) for the analyses of focal alterations in thisstudy.Joel GreshockBarbara L. WeberDiscovery and Translational Medicine,GlaxoSmithKline,King of Prussia, PennsylvaniaBin FengElena IvanovaIlana PernaAlexei ProtopopovLynda ChinCenter for Applied Cancer Science,the Belfer Institute forInnovative Cancer Science,Boston, MassachusettsCristina NogueiraLynda ChinDepartment of Medical Oncology,Dana-Farber Cancer Institute,Boston, MassachusettsLynda ChinDepartment of Dermatology,Harvard Medical School,Boston, MassachusettsKatherine NathansonAbramson Family CancerResearch Institute,University of Pennsylvania,Philadelphia, Pennsylvania
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    2
    References
    1
    Citations
    NaN
    KQI
    []