As the wide popularization of online social networks, online users are not content only with keeping online friendship with social friends in real life any more. They hope the system designers can help them exploring new friends with common interest. However, the large amount of online users and their diverse and dynamic interests possess great challenges to support such a novel feature in online social networks. In this paper, by leveraging interest-based features, we design a general friend recommendation framework, which can characterize user interest in two dimensions: context (location, time) and content, as well as combining domain knowledge to improve recommending quality. We also design a potential friend recommender system in a real online social network of biology field to show the effectiveness of our proposed framework.

As hundreds of millions of users access online social communities at daily or even real-time basis, large amounts of user data are continuously generated. The fast-growing user-generated content poses great challenges on online social network system design for efficient content management and delivery. For instance, in content-centric online social network, all contents are organized in a temporal order which makes it very time consuming for users to browse all these contents to locate what they really like. By conducting a comprehensive study of online user activities, including content, social, and time characteristics, this paper try to accurately characterize user interest and user context, with the end goal of more efficient content management and real-time content delivery in online social network systems. The detail analysis of user activities is conducted on the real data collected from a popular online social community among Chinese universities with over 63,000 users to demonstrate the advantage and effectiveness of extracting user interest and context characteristics and applying them in designing more efficient content management systems.

DNA transposon piggyBac (PB) is a newly established mutagen for large-scale mutagenesis in mice. We have designed and implemented an integrated database system called PBmice (PB Mutagenesis Information CEnter) for storing, retrieving and displaying the information derived from PB insertions (INSERTs) in the mouse genome. This system is centered on INSERTs with information including their genomic locations and flanking genomic sequences, the expression levels of the hit genes, and the expression patterns of the trapped genes if a trapping vector was used. It also archives mouse phenotyping data linked to INSERTs, and allows users to conduct quick and advanced searches for genotypic and phenotypic information relevant to a particular or a set of INSERT(s). Sequence-based information can be cross-referenced with other genomic databases such as Ensembl, BLAST and GBrowse tools used in PBmice offer enhanced search and display for additional information relevant to INSERTs. The total number and genomic distribution of PB INSERTs, as well as the availability of each PB insertional LINE can also be viewed with user-friendly interfaces. PBmice is freely available at http://www.idmshanghai.cn/PBmice or http://www.scbit.org/PBmice/.