Genomics is becoming a data-intensive science, and an increasing number of laboratories are generating data which swamps storage in traditional paper-and-ink notebooks. Capturing the data flow requires large systems with multiple applications manipulating the same or similar data. Large systems often have conflicting requirements for data representation. Consistency across applications is a prime consideration, and appropriate data representation is an important issue in developing practical systems for molecular biologists. Graphs are a natural representation for describing genome data, while objects are good for modeling the behavior necessary for laboratory applications. We present a method for translating graph descriptions of genome data into objects using objects as views on graphs. Graph representations describe genome concepts while objects capture individual views for application development insuring consistency across genome applications.