Many programs of bioinformatics provide biochemists and biologists retrieve and analysis services of gene and protein databases. These services access databases for each arrival of user's request, which takes a long time, increases server's overload and response time. In this paper, two data sharing models are presented to reduce database retrieve for sequence alignments. (1) Grouping model is proposed to share database access between many requests arrived in a certain period of time by utilizing database access patterns in bioinformatics. (2) Carpool model is also proposed to reduce response time as well as to increase system throughput by servicing new arriving request immediately together with the previous on going requests to share database accesses without waiting time. Simulation results show that two data sharing models can reduce the number of database access by sharing it among many requests and increase system throughput.