LIMS and the art of MS proteomics
No longer just for the pharmaceutical industry, LIMS has a lot to offer to MS proteomics.
A laboratory information management system (LIMS) is a software package that can manage laboratory samples, users, instruments, and data. The most extensive type of LIMS—enterprise LIMS—is designed to link laboratory systems to an organization’s business systems, such as personnel and equipment management, inventory, and accounting. According to Mike Sanders of GenoLogics Life Sciences Software, Inc., in practice, many people just want LIMS to track data and samples and to optimize their laboratory’s operation.
This review will focus on the more basic LIMS rather than the enterprise type. The organizations invited to participate were identified by three channels. First, commercial and freely available LIMS packages specifically tailored to, or at least mentioning, MS were identified through Internet research. Second, because major vendors can customize general LIMS packages for MS-based proteomics (and may already have done so), the seven most commonly cited vendors for general LIMS packages, as listed in Scientific Computing’s 2007 LIMS Market Watch Report, were contacted (1). Third, proteomics researchers at the Laboratory of Neurotoxicology at the U.S. National Institute of Mental Health made suggestions based on their recent experience of choosing a LIMS. Not all of the organizations that were contacted responded. The commercial LIMS are listed in Table 1 and the free LIMS in Table 2. Those systems produced by companies exclusively for controlling their own brand of instruments were omitted. Finally, many more LIMS products are tailored to gel-electrophoresis- and microarray-based proteomics, but for brevity, these products are not covered here.
History
LIMS software evolved as a product of laboratory automation, and the pharmaceutical industry developed the first in-house laboratory management software in the late 1960s (2, 3). Although some disagreement exists as to when the first commercial LIMS was offered, Beckman Coulter’s Laboratory Automation Operation introduced a commercial software package for laboratory management in 1979, and the term LIMS was coined when PerkinElmer launched its LIMS/2000 in 1981 (4, 5).
Key features
So, what do you need to know before choosing a LIMS for your laboratory? For proteomics research, most LIMS features can be divided into five general areas: requirements and capabilities, data, integration, flexibility, and security and compliance.
Requirements and capabilities. One of the first considerations is whether the LIMS comes ready to use “out of the box”, is custom built, or is an out-of-the-box product with additional customization (6). Some researchers can find a package that already has all of the elements that their laboratory needs, but others may need to add special features. Clearly, the implementation time will be longer for any system requiring customization. “You want to make sure you know what you need,” says Leonard Foster of the University of British Columbia (Canada). “But you also need to make sure that you’ve got the know-how in your lab to make use of [the LIMS].”
Other requirements include where the program is installed and where it runs. There are different advantages to installing and running software on each user’s computer (the fat or thick client model), storing the software and doing all of the processing on a central server (the thin client model), and hybrids of the two. Keith O’Leary of LabVantage Solutions, Inc., says that thin client systems are ideal because you don’t have to run any programs on your own computer. “Many customers just won’t let you download ActivX or any other kind of plug-in,” he says. All of the LIMS listed in this review can run on a PC except for CPAS, ProteinCenter, and SORCERER 2. Now, most LIMS products offer a web-based interface, so the system can be accessed from any location that has an Internet connection.
Data. Another important category of LIMS features has to do with how the software acquires, presents, stores, and analyzes data. For proteomics data, two different standard XML formats, mzData and mzXML, were previously available. A new standard format, mzML, has been developed to replace the other two, but many users will need a system that is compatible with old and new data formats. To work with images and sequences, a LIMS needs to be capable of handling binary large objects (BLOBs) and character-based large objects (CLOBs), respectively. Other data formats such as comma-separated values (.csv), Microsoft Word (.doc), PDF, XLS, and XML are useful for importing the data into other programs.
Searching for samples, proteins, and user-defined fields is generally done via a user interface or query language. “Once you’ve collected and gathered all of [these] data, the question would be: how do you actually query that data and drill through to getting subsets of data?” says O’Leary. All of the LIMS except LIPAGE can search the database without a query language. And once researchers find the data subsets, they will want to see relationships among sample characteristics, methods, and results. Foster says that for proteomics, the most important LIMS features are organizing data and recording specific parameters about the data so you can go back and mine them later. Rob Ewing of Case Western Reserve University agrees. “We certainly find that being able to go back to the data in a systematic way and reanalyze the data is probably the most important [benefit],” he says. For example, the departure of a group member can cause problems with data continuity. Traditionally, colleagues would pore over the former member’s laboratory notebook to try to figure out the experimental conditions and results, but LIMS eliminates this problem. “It allows better data management, and it provides a mechanism for secure storage of data that no longer relies on the individual lab members,” Foster says.
Integration. A LIMS can have one- or two-way communication with a variety of instruments to receive the data or even control the instruments. For proteomics, it is crucial that a LIMS can interface with third-party software and databases (7). James DeGreef of GenoLogics sees a trend toward using multiple protein search engines to analyze and validate data. Some users will also want to compare data from multiple search engines. Foster says that for his lab, the most important feature is “the ability to control third-party software . . . basically to build a data pipeline beyond what the LIMS itself provides.”
Flexibility. To meet the future needs of the laboratory, systems must be able to interface with new instruments as they are added and with new types of instruments. All of the LIMS software packages listed in Tables 1 and 2 are expandable and can support new technologies.
Often, users need to make changes to the workflow. Part of a system’s flexibility is how these changes are made—through programming or menus. Similarly, systems may or may not allow users to add to the database entities such as new tables that track data about the supplier of a particular reagent; systems that allow such additions may effect these changes through menus or programming. The less programming required after implementation, the more user-friendly a system should be. Ewing thinks that the most important LIMS feature is usability. “I think people need to find systems which are in line with their experience and their skill sets. You don’t want to have a system which is unusable by your staff,” he says. Another element of LIMS flexibility is customized data output. Except for ProteinCenter, all of the LIMS in this review allow the user to create customized reports.
Security and compliance. Varying levels of privilege and access help to prevent accidental modification or loss of data, but they can also enable external connections. Terry Smallmon of LabVantage says, “You’ve got to expand beyond the walls of the laboratory to the outer world, so now you need a bunch of different types of security.” Some laboratories give external collaborators or customers limited access to the LIMS, allowing them to see just their own data. All of the LIMS in Table 1 and PRIME and Proteios Software Environment in Table 2 provide security via different levels of user access. All of the LIMS except for ProteinCenter and SORCERER 2 in Table 1 and LIPAGE and ms_lims in Table 2 offer sample tracking and/or an audit trail. Compliance is more important in industry where, for example, experiments must be documented in a way that is acceptable to the U.S. Food and Drug Administration or other regulatory agencies when filing for drug approval.
Commercial vs open source
Two different camps have emerged for LIMS—commercial and open source. In open-source software, the source code is accessible, so the user can change and improve the software. These types of computer programs are usually available free of charge. But as some people are quick to point out, you must take into account the total cost of ownership because in the long run, you might be dedicating more of your bioinformatics support to the software than originally planned. Some researchers see a trend among MS proteomics groups toward using more open-source LIMS, but they caution that, in those cases, the user depends on the software creators to stay current and make the program compatible with new types of third-party analyses as they are created. Moreover, DeGreef says that you have to make sure that the modifications you make to the underlying software will integrate with updates by the original developers. But he also notes that open-source algorithms have been particularly useful in conjunction with LIMS software. “I think that’s where open source can shine really well,” he says. None of the commercial LIMS (Table 1) are open source except for SORCERER 2, which is partially open source, and CPAS, which is completely open source. All of the free LIMS (Table 2) are completely open source.
Advantages
The foremost advantage of using a LIMS is that the automation of experiments and data analysis can dramatically increase a laboratory’s productivity. Foster says, “I guess the thing that has been most surprising was how easy it was to set up a data pipeline and how much time that saved us.” Accessibility to data is significantly improved, particularly if a web-based interface allows access from remote locations. (All of the LIMS packages reviewed here have a web-based interface except Biotracker and ms_lims.) In addition, traditional laboratory notebooks are not compatible with a multiuser, multitask environment, so an electronic means of storing and sharing data is an attractive option (8). Last but not least, because a LIMS records experimental details and data in an electronic format, it has the potential to avoid a significant amount of human error and to reduce the consumption of paper.
Other considerations
As can be seen from Tables 1 and 2, costs vary widely among LIMS providers. Researchers must be sure to include estimates for implementation, training, management, and maintenance to get an accurate picture of the total cost of ownership. They must also take into account the type of licensing strategy a provider offers. For example, under some licenses, providers charge for a certain number of concurrent users; under others, they charge by the number of named users. Moreover, implementation may prove challenging. Foster says, “There can either be a short or very long adoption time where people begin to actually start using the system. So it’s taken some pressure—and a lot of pressure in some cases—to get some people to buy into using the system.”
Future prospects
Initially, some researchers had high expectations for proteomics, thinking that it might even eclipse genomics. “The proteomics world promised a new dimension. It was going to be genomics on steroids,” says Smallmon. But proteomics has been slow to fulfill expectations. In fact, several LIMS systems developed for MS-based proteomics (e.g., ProtLIMS, ProteomIQ) are no longer updated or available. Instead of pure proteomics, Sanders sees a trend toward interdisciplinary research. “People are looking across the sciences more and more,” he says. “People are wanting . . . to integrate data in more of a systems-biology perspective.” Some of the LIMS products already on the market are designed to accommodate multiple domains of science, and more are likely to follow.
Christine Piggee is an associate editor of Analytical Chemistry.
References
- Scientific Computing. LIMS Market Watch Report: Executive Summary; Oct 2007. www.scientificcomputing.com/pdfs/2007/LIMSMW2007/ExecSummary.pdf.
- Gibbon, G. A. Lab. Autom. Inform. Manag. 1996, 32, 1–5.
- Peden, J. Gearing Up LIMS for the Challenge of Burgeoning Bioinformatics. eLab, Jan 12, 2001; http://snipurl.com/2axnc.
- Hall, P. Acquisitions and Mergers Take Centre Stage. Scientific Computing World, Nov/Dec 2003; www.scientific-computing.com/features/feature.php?feature_id=103.
- Association for Laboratory Automation. Creative Combustion: A History of the Association for Laboratory Automation; www.labautomation.org/about/timeline/combustion.html.
- Avery, G.; McGee, C.; Falk, S. Anal. Chem. 2000, 72 (1), 57 A–62 A; http://pubs.acs.org/hotartcl/ac/00/jan/prodreview.html.
- Smallmon, T. R.; Ganjei, J. K. The Benefits of a LIMS in Proteomics. LIMSLetter, Winter 2004, 1–5, 8.
- Haquin, S.; et al. Data Management in Structural Genomics: An Overview. In Structural Proteomics: High-Throughput Methods; Kobe, B., Guss, M., Huber, T., Eds.; Methods in Molecular Biology; Humana Press: Totowa, NJ, 2008; Vol. 426, pp 49–79.
Advertisements
Return to Top
Copyright © American Chemical Society, 1155 Sixteenth Street N.W., Washington, DC 20036
