Review Paper N.M. Luscombe, D. Greenbaum, Review M. Gerstein Department of Molecular Biophysics and Biochemistry What is bioinformatics? An Yale University New Haven, USA introduction and overview Abstract: A flood of data means that many of the challenges in biology are now challenges in computing. Bioinformatics, the application of computational techniques to analyse the information associated with biomolecules on a large-scale, has now firmly established itself as a discipline in molecular biology, and encompasses a wide range of subject areas from structural biology, genomics to gene expression studies. In this review we provide an introduction and overview of the current state of the field. We discuss the main principles that underpin bioinformatics analyses, look at the types of biological information and databases that are commonly used, and finally examine some of the studies that are being conducted, particularly with reference to transcription regulatory systems. Introduction Bioinformatics - a definition1 Biological data are being produced (Molecular) bio – informatics: bioinformatics is conceptualising biology in at a phenomenal rate [1]. For terms of molecules (in the sense of physical chemistry) and applying example as of August 2000, the "informatics techniques" (derived from disciplines such as applied maths, GenBank repository of nucleic acid computer science and statistics) to understand and organise the information sequences contained 8,214,000 associated with these molecules, on a large scale. In short, bioinformatics entries [2] and the SWISS-PROT is a management information system for molecular biology and has many database of protein sequences practical applications. contained 88,166 [3]. On average, these databases are doubling in 1 As submitted to the Oxford English Dictionary size every 15 months [2]. In addition, since the publication of the H. As a result of this surge in data, that life itself is an information influenzae genome [4], complete computers have become indispensable technology; an organism’s physiology sequences for over 40 organisms to biological research. Such an approach is largely determined by its genes, which have been released, ranging from is ideal because of the ease with which at its most basic can be viewed as 450 genes to over 100,000. Add to computers can handle large quantities digital information. At the same time, this the data from the myriad of of data and probe the complex dynam- there have been major advances in the related projects that study gene ics observed in nature. Bioinformatics, technologies that supply the initial data; expression, determine the protein the subject of the current review, is Anthony Kerlavage of Celera recently structures encoded by the genes, often defined as the application of cited that an experimental laboratory and detail how these products inter- computational techniques to understand can produce over 100 gigabytes of act with one another, and we can and organise the information associated data a day with ease [5]. This incredible begin to imagine the enormous with biological macromolecules. This processing power has been matched quantity and variety of information uexpected union between the two by developments in computer technol- that is being produced. subjects is largely attributed to the fact ogy; the most important areas of Yearbook of Medical Informatics 2001 83 Review Paper improvements have been in the CPU, understanding , large-scale and be listed. We also give approximate disk storage and Internet, allowing practical applications . Specifically, we values describing the sizes of data being faster computations, better data stor- discuss the range of data that are discussed. age and revolutionalised the methods currently being examined, the databases for accessing and exchanging data. into which they are organised, the types We start with an overview of the of analyses that are being conducted sources of information: these may Aims of bioinformatics using transcription regulatory systems be divided into raw DNA sequences, The aims of bioinformatics are three- as an example, and finally some of the protein sequences, macromolecular fold. First, at its simplest bioinformatics major practical applications of structures, genome sequences, and organises data in a way that allows bioinformatics. other whole genome data. Raw DNA researchers to access existing infor- sequences are strings of the four base- mation and to submit new entries as letters comprising genes, each typically they are produced, eg the Protein Data “…the INFORMATION 1,000 bases long. The GenBank Bank for 3D macromolecular struc- associated with these repository of nucleic acid sequences tures [6,7]. While data-curation is an molecules…” currently holds a total of 9.5 billion essential task, the information stored bases in 8.2 million entries (all database in these databases is essentially use- Table 1 lists the types of data that are figures as of August 2000). At the next le ss until analysed. Thus the purpose of analysed in bioinformatics and the range level are protein sequences comprising bioinformatics extends much further. of topics that we consider to fall within strings of 20 amino acid-letters. At The second aim is to develop tools and the field. Here we take a broad view and present there are about 300,000 known resources that aid in the analysis of include subjects that may not normally protein sequences, with a typical data. For example, having sequenced a particular protein, it is of interest to Table 1. Sources of data used in bioinformatics, the quantity of each type of data that is currently compare it with previously characte- (August 2000) available, and bioinformatics subject areas that utilise this data. rised sequences. This needs more than just a simple text-based search and Data source Data size Bioinformatics topics Raw DNA sequence 8.2 million sequences Separating coding and non-coding regions programs such as FASTA [8] and (9.5 billion bases) Identification of introns and exons PSI-BLAST [9] must consider what Gene product prediction comprises a biologically significant Forensic analysis match. Development of such resources Protein sequence 300,000 sequences Sequence comparison algorithms (~300 amino acids Multiple sequence alignments algorithms dictates expertise in computational each) Identification of conserved sequence motifs theory as well as a thorough under- Macromolecular 13,000 structures Secondary, tertiary structure prediction standing of biology. The third aim is to structure (~1,000 atomic 3D structural alignment algorithms use these tools to analyse the data and coordinates each) Protein geometry measurements Surface and volume shape calculations interpret the results in a biologically Intermolecular interactions meaningful manner. Traditionally, Molecular simulations biological studies examined individual (force-field calculations, systems in detail, and frequently molecular movements, compared them with a few that are docking predictions) related. In bioinformatics, we can now Genomes 40 complete genomes Characterisation of repeats (1.6 million – Structural assignments to genes conduct global analyses of all the 3 billion bases each) Phylogenetic analysis available data with the aim of un- Genomic-scale censuses (characterisation of protein content, metabolic pathways) covering common principles that apply Linkage analysis relating specific genes to diseases across many systems and highlight Gene expression largest: ~20 time Correlating expression patterns novel features. point measurements Mapping expression data to sequence, structural and for ~6,000 genes biochemical data In this review, we provide an intro- duction to bioinformatics. We focus on Other data the first and third aims just described, Literature 11 million citations Digital libraries for automated bibliographical searches with particular reference to the key- Knowledge databases of data from literature words underlined in the definition: infor- Metabolic pathways Pathway simulations mation,informatics, organisation, 84 Yearbook of Medical Informatics 2001 Review Paper bacterial protein containing approxi- sequences. While more biological infor- relationship between the two proteins mately 300 amino acids. Macromo- mation can be derived from a single is remote [17, 18]. Among homologues, lecular structural data represents a structure than a protein sequence, the it is useful to distinguish between more complex form of information. lack of depth in the latter is remedied orthologues, proteins in different There are currently 13,000 entries in by analysing larger quantities of data. species that have evolved from a the Protein Data Bank, PDB, most common ancestral gene, and of which are protein structures. A paralogues, proteins that are related by typical PDB file for a medium-sized “… ORGANISE the informa- gene duplication within a genome [19]. protein contains the xyz coordinates tion on a LARGE SCALE …” Normally, orthologues retain the same of approximately 2,000 atoms. function while paralogues evolve Redundancy and multiplicity of data distinct, but related functions [20]. Scientific euphoria has recently A concept that underpins most centred on whole genome sequencing. research methods in bioinformatics is An important concept that arises As with the raw DNA sequences, that much of this data can be grouped from these observations is that of a genomes consist of strings of base- together based on biologically meaning- finite “parts list” for