The rule of thumb is to postpone any genome-wide analysis until the genome paper is published. On the other hand, using deposited data to perform a vertical analysis (analysis of a single gene family across several species) is usually accepted. However, the best idea is to ask the sequencing lab. You'd be surprised how nice people are usually working there ;)

As alluded to by freesci, there is a standard policy referred to as the Ft. Lauderdale agreement, which is detailed here: http://www.genome.gov/10506537

Some feel this agreement is restrictive and have proposed other licenses for genome data (e.g. the Batavia Open Genomic Data Licence) and BGI has released some data into the public domain directly.

However, in the absence of explicit wording otherwise, assume the data are covered by the Ft. Lauderdale agreement and contact the group who did the sequencing to coordinate publication.