If a company obtains a patent on an important gene, no one can else can do research on the gene and subsequently use it, and thus important (medical) advances may be delayed.
Patent law permits non-commercial research on patented subject matter, so pure research by academic institutions is not affected by the existence of patents.

Patent law has a "research exemption" rule, whose goal is precisely not to block fundamental research. Any breeder or any company has a free access to patented animals or genes and may experiment in order to develop a new application of the gene or to perfect a protected method.
But as its name suggests, research exemption is only valid within the confines of research.

As soon as the breeder or the company develops an invention closely related to the patented one (for ex. a new application of a patented gene) and wants to commercialize it, he may be obliged to get the authorization of the patentee. The latter is not obliged to grant a license and may prevent the second invention from being commercialized.

However, in relation to commercial research it does not necessarily follow that others are irrevocably blocked as it will often be possible to negotiate a licence under the patent – or challenge its validity. The pharmaceutical industry is opposed to any attempt to impose compulsory licensing as that would undermine the incentives that the patent system provides. A better way is to rely on voluntary licence agreements as we always have done.

Finally, it is always open to third parties to obtain 'dependent patents' - that is to say, patent a new use for an already patented gene. The original finder of the gene could not then commercialise the new use without a licence under the dependent patent. This situation, which tends to stimulate cross-licensing, is inherent in the patent system. It is frequently encountered with pharmaceuticals and is in principle no different in the genomics field.