W. John MacMullen

Information problems in molecular biology and bioinformatics

In this article we provide an overview of opportunities for research and practice in the domain of molecular biology by information and library scientists. We introduce the changing role of data and information in molecular biology, and how molecular biology is evolving from a technique- and technology-driven science to an information-driven science. We then describe the high-level objectives of molecular biology and some broad classes of problems from an information perspective. We illustrate the high-level objectives with examples of specific tasks performed by biologists. Finally, we provide some programmatic direction for information and library science research streams and insertion points.

Characteristics of bioinformatics employment advertisements

Current trends in the bioinformatics job market were assessed using a sample of 1,996 online employment advertisements from the 6-year period of January 2003 through December 2008. Job postings were classified by employer type, job role, and location, and a content analysis of the text of a subset of 404 of the posts was performed to identify detailed characteristics associated with masters degree level positions, such as educational requirements and preferred scientific and technical skills. Consistent with previous studies, academic institutions, corporations, and research institutes are the primary employers of bioinformaticists. In the U.S. only three states, California, Maryland, and Massachusetts, provided the majority of opportunities. Graduates from all levels of education are needed in the field, although those with a bachelors degree in computer science or a Ph.D. in biology or bioinformatics are especially in demand. Perl programming is the most frequently requested skill across advertised positions, and experience using bioinformatics software tools (such as BLAST, CLUSTAL, and HMMER) is the bioinformatics skill in greatest demand. A small number of positions specifically requesting librarians and LIS-trained bioinformaticists was observed, but the significant bioinformatics activities in biomedical libraries over the same time period are not reflected in this snapshot of the bioinformatics workforce.

Inter-database annotation linkages in model organism databases

Inter-database linkage via annotations is one approach to the integration of knowledge in the biomedical domain, which is often fragmented by specialization. This pilot study examined annotations in ten model organism databases and the Gene Ontology (GO) to assess explicit and implicit linkages between organisms. The databases had similar annotation processes, content, and knowledge, with some variation due to organizational objectives and technology infrastructure. While all databases had the potential to be linked to all others via GO, only some databases had non-GO links to others. This may be due in part to a lack of biologically significant relationships among some of the organisms, or that putative relationships between them that exist in GO have not yet been explored.

Facets and measures of gene ontology annotation quality in model organism databases

Model organism databases are important repositories of data and information for biomedical research, but are useful to scientists only if the information they contain meets certain levels of quality. This methodology paper describes six facets of information quality applicable to Gene Ontology (GO) annotations in model organism databases, and defines corresponding metrics to be used in measuring the quality of annotations made by one or more human database curators. The defined facets and measures of annotation quality are: consistency, reliability, specificity, completeness, accuracy, and validity. Contextual factors, and factors affecting internal and external validity, are also discussed.

Quantifying literature citations, index terms, and Gene Ontology annotations in the Saccharomyces Genome Database to assess results-set clustering utility

A set of 37,325 unique literature citations was identified from 120,078 literature-based annotations in the Saccharomyces Genome Database (SGD). The citations, gene products, and related Gene Ontology (GO) annotations were analyzed to quantify unique articles, journals, genes, and to rank by publication year, language, and GO term frequency. GO terms, MeSH indexing terms, MeSH Journal Descriptors, and SGD Literature Topics were quantified and analyzed to assess their potential utility for results set clustering. Results: Bradford’s Law of Scattering was shown to hold for the citations, journals, gene products, and GO annotations. Only the MeSH terms and article title/abstract pairs had significant numbers of term co-occurrence. Multiple term types may be useful for faceted searching and clustered results set browsing if the strengths of each are leveraged.

Searching for uses and users in Gene Ontology research

Introduction: Despite millions of dollars in investment over the past decade in the creation and maintenance of the Gene Ontology (GO), little is known about how (or even if) its intended end users – biomedical researchers – actually employ the ontology and its related databases and interfaces in their work. This project is a preliminary investigation of what evidence exists in the literature of specific uses of GO by researchers, and of use cases proposed for researchers by system designers. This work will help inform future in-depth studies of the specific information needs and research questions that researchers might use GO and other similar knowledge structures to address. It also provides to library and information science researchers and practitioners some insight into the quantity, sources, and breadth of publications about GO that exist.

Evolving curricula in LIS-focused bioinformatics programs

Introduction At the 2002 ASIST Bartlett & Toms, 2005), rather than taking an educational perspective.

The effects of context on data quality in biomedical data reuse

Introduction The collection of huge stores of electronically-formatted information and advances in information processing technologies has led to dramatic changes in the conduct of biomedical science. In biology, a paradigm shift is underway due to an unprecedented flood of data, the emergence of shared research repositories, and advances in the application of data mining algorithms. As a result, the traditional model of scientific discovery of “formulate hypothesis, conduct experiment, evaluate results” is being replaced with “collect and store data, mine for new hypotheses, confirm with data or supplemental experiment” (Han et al. 2002). In clinical medicine, similar developments have made possible a variety of secondary applications of extant clinical data, including physician decision support, outcomes assessment, document retrieval, and clinician performance evaluation. The quality, or usefulness, of existing data for secondary uses has thus far been approached with a focus on issues of technical access and the mathematical format of data. Largely ignored in these efforts are the effects on quality introduced when data captured for one purpose is reused for another. This panel brings together four researchers with on going studies in the biomedical domain focused on the effects of context on the quality of data for secondary uses. Drawing from empirical evidence the topics to be addressed include; 1) the effects of context on data reuse, 2) anticipating, identifying and accounting for these effects, and 3) the future implications of large-scale data reuse for information professionals in the biomedical domain.

Managing scientific data for long‐term access and use

Preservation of data for long-term use will require data management strategies that include curation and preservation planning and implementation. While data management and curatorial activities have been an integral part of some scientific domains for years (see for example, high energy particle physics), these are new concepts in other areas of science. Concepts such as provenance, representation for re-use, and work-flow capture are rarely understood, let alone addressed. By bringing together theories and best practices from archives, museum studies, and library and information science (LIS), it is possible to address these problems. on current research into scientific data management problems, this panel will consider questions about sharing and re-use of data, curation and preservation, and the intersection of scientific production and scholarly communication. Our research explores information work and problems across a range of scientific areas in the life and physical sciences, including genomics, neuroscience, ecology, and earth science. As more scientific work products are shifted to open or shared data collections (including archives, repositories and databases), we will need to understand how these systems are implemented and used to support collaboration and discovery, as well as scholarly and scientific communication.