The TIMTAM ("Target Information Management Tracking And More") system implements a knowledgebase supporting the management, tracking, and analysis of experimental data associated with protein crystallography projects. It responds to a crucial need within the structural biology community, allowing researchers to effectively manage the volume and diversity of information generated by high throughput techniques.
In general terms, TIMTAM is referred to as a Laboratory Information Management System (LIMS). In addition to storing results, it models the work processes of the laboratory, including users, projects, samples, experiments, and protocols. TIMTAM is employed following the target selection phase of high-throughput crystallography projects and models the cloning, expression, purification, and crystallisation steps that precede structure determination from diffraction quality crystals.
The first step in using TIMTAM is the creation of a new project by a managing scientist. This scientist is the owner of the project and is able to specify read-only or owner access for other laboratory users in their team. This system of permissions is designed to protect the confidentiality of experimental results and intellectual property prior to publishing. Several steps are then defined corresponding to the experimental workflow of the laboratory; typical examples include "Cloning" and "Small Scale Expression and Purification", although these can be customised entirely on a per-project basis. Each step is defined by an experiment type composing a set of parameters for recording experiment results. These parameters may contain numerical values, textual description, images, or files. For example, the "Cloning" step might include an analysis named "PCR/Caliper" with parameters for size, concentration, molarity, and gel images. Figure 1 below shows the user interface for this example. Definitions for experiment types and parameters can be added or changed during the course of a project, providing a significant degree of flexibility and system adaptability to suit scientists' requirements.
Following definition of a project, information is entered about targets and their associated constructs. Target information can include sequences for multiple organisms (such as human and mouse), external identifiers linking to the Entrez and Ensembl databases, and notes made by scientists. Each target is also assigned a "person in charge" to reflect the organisational structure of the project. Constructs are created by specifying a subsequence of the parent target and any forward and reverse primers used. Data entry for targets and constructs is assisted by: automatic conversion between protein and DNA sequences; automatic calculation of values such as molecular weight, pI, and crystal scores; and retrieval of information from public databases.
Scientists submit experimental results for each construct by completing a simple form specifying parameter values for the analysis performed. Figure 2 below illustrates an experiment with two images (an SDS-PAGE gel and an SEC chart) and several analysis parameters (Yield, Purity, MW, Homegeneity, Stability). In this case, the tick in the top right hand corner marks this experiment as successful and the scientist has provided a comment at the bottom explaining the outcome.
The header in Figure 2 also includes references to the tube ("A1") and microplate ("P020") containing the sample being tested. These unique identifiers assist with record keeping within the laboratory by matching physical samples to database entries. Users are able to record the contents of microplates through an interface that allows them to drag-and-drop the combination of construct and vector contained in each tube. Figure 3 illustrates an example of this interface.
Each construct is tracked by the system as it progresses through the steps of the project. The user interface for monitoring construct progress is shown in Figure 4 below. Individual analyses are displayed as a tick, cross, or question mark to respectively indicate the success, failure, or unreviewed outcome of an experiment.
Each step comprises a set of experiments, and these experiments can be sub-categorised according to the vectors used within the experiment. Initially, the table (as shown in Figure 4) displays a single icon which indicates the most successful outcome among all the experiments for each step. However, step columns can be expanded to display more precise results. For example, the "Small Scale Exp & Purif" column can be expanded to show outcomes for experiments goruped by vector. The "Cloning" column has beed expanded to display results for the two vectors (pMCSG7 and pMCSG19C) used on this project. The progress view is useful for both scientists and managers overseeing the project.
An important feature in the progress view is the "E-value", which measures similarity between a construct and sequences published in PDB/ohPDB. It is updated weekly by retrieving information from PDB and ohPDB and using BLAST to measure similarities. If the E-value is below a certain threshold (1.0e-10), an email is sent to the target owner to notify them that a homologous target has been solved.
TIMTAM is implemented as a Web application, meaning that users can access it through a Web browser without having to install any special software. Unauthorised access is restricted by a secure login and password system. The Web server component is based on the Ruby on Rails application framework, MySQL relational database, and the BioRuby and BioPerl code libraries, all of which are freely available and run on all major operating systems.
TIMTAM was designed to resolve the complexity of managing data generated by high-throughput protein crystallography projects and arguably achieves this by providing a powerful but easy-to-use database application. It is currently being used by the Structural Genomics group at The University of Queensland, and it is hoped that systems such as TIMTAM will become more widespread in the future as their benefits are realised by protein crystallographers.