Automatic lane and band detection in Image Lab 5.2.1. Credit: Bio-Rad Laboratories)

The specter of the irreproducibility crisis looms across the scientific landscape (Nature, Challenges in Irreproducible Research, 2016; Science, Journals Reunite for Reproducibility, 2014). No study or technique is immune from its grasp, certainly not the 40-year old western blot. At the turn of this decade, the United States Office of Research Integrity estimated that 70 percent of images (not just of western blots) were deliberately manipulated. To reduce such occurrences, and to encourage reproducibility and transparency, many journals/institutes have begun placing more stringent guidelines on the analysis of protein gels and immunoblots. Consequently, several digital imaging systems that enable scientists to adhere to the above guidelines have emerged in the market. These digital imagers are equipped with cutting-edge image analysis software that can facilitate higher reproducibility of data through a variety of automated features. However, the available software can differ in their degree of automation and ease of use and functionality, which then raises another question for researchers: “What to consider when investing in a digital imaging system with dedicated analysis software”.

Pre-defined workflows that can improve reproducibility

One of the most attractive features of a dedicated image acquisition and analysis software is the availability of pre-defined application protocols that are already programmed with optimized light sources and detection filters. These mostly-automated routines present a combination of settings for imaging, analyzing and reporting that can be run as a single, 100 percent repeatable workflow. They often require no training and provide a quick, straightforward approach to obtaining an image, without the hassle of a long-drawn-out, trial-and-error optimization process. Such protocols are particularly useful when analyzing a large quantity of blots.

For example, in the Image Lab software system optimization initiates at setup and all following images are obtained with predefined protocols, rendering them reproducible and free of imaging artifacts. Additionally, the lane detection and crop size settings can be saved and applied to all images to be acquired. Images are focused in the same way, regardless of zoom level or sample position. Auto-focus and auto-exposure features effectively minimize human intervention during image capture, saving time and effort as well as ensuring reproducibility.

Data security features that preserve the original data

The Journal of Biological Chemistry has some of the most detailed and rigorous guidelines for western blotting data. To ensure data integrity, journals are increasingly requiring authors to provide the original (raw) immunoblot data before modification in Photoshop and Image Lab (see PLOS One Submission Guidelines).

New imaging systems can ensure the preservation of original blot images from alteration. These features will help scientists instill confidence in their data and protect against claims of data manipulation. Furthermore, such features can help pharma and biotech researchers more readily meet the FDA's regulations for current Good Manufacturing Practice environments and comply with rules for secure handling of electronic records.

Some imaging software enables users to protect their data files, prevent deletion of files and limit where data can be saved. In the Image Lab Software, for example, this is made possible through several options, including access controls and authority checks via user identification codes, electronic record security via protected directories, and date and time-stamped audit trails. 

Advanced data analysis features that can deliver trustworthy results

JBC guidelines prefer “normalization of signal intensity to total protein loading” over the use of housekeeping proteins. This normalization technique, called total protein normalization (TPN), relies on the entire complement of proteins in a lane rather than on individual housekeeping proteins, which have now been shown to demonstrate some variability in expression. In TPN, the band of interest is compared to the abundance of total protein in each lane (based on a stain such as Ponceau S), and the ratio of band intensity to total protein is obtained.

Novel image analysis software can automatically detect protein lanes and bands in the gel, enabling users to carry out the TPN technique for western blot normalization, with minimal human error. Furthermore, the latest software versions are compatible with Stain-Free technology for TPN, which eliminates the use of additional staining reagents and steps to visualize proteins, while also providing higher sensitivity than anionic stains.

Automated capabilities that make the entire western blot workflow more efficient and reproducible have been integrated into dedicated imaging solutions. Automation of steps like background subtraction can aid in removing human error from the analysis process. For instance, digital imagers with advanced software are now able to automatically conduct localized (rolling-disc) background subtraction around each band, instead of using a single background value common to all lanes. This capability does not require any user intervention, thus removing human error from the analysis process. In some software, it is even possible to preview the image before acquisition, allowing the user to highlight a region of interest that needs to be acquired with the clearest signal.

The right image analysis software can play a big role in resolving reproducibility problems and providing researchers a largely automated and hands-off approach to data analysis. With versatile analytical functions, dependable preservation of original data, and access to pre-programmed protocols that eliminate the need for specialized knowledge on optics, researchers can have greater confidence in the quantitation of their western blot images.