Rigor and Reproducibility
Eight steps to Rigorous and Reproducible Experiments in Biomolecular Research at UNC:
1. If using a core facility, consult with the core staff in the planning stage. Consult with a statistician if you need help developing a Power Analysis to assure that your results will be adequately powered.
2. Design your experiment with sufficient controls (rigor) and replicates (reproducibility).
3. Assure that ALL of your reagents (antibodies, cell lines, mice) are fully validated (see below).
4. Have a clear and detailed protocol (SOP) and data analysis plan. Assure that the protocol is strictly followed or that any deviation is well documented.
5. Assure that the staff or students performing the experiment are well trained and understand each step and the importance of performing them precisely.
6. Use only well-maintained instrumentation, preferably maintained and operated in a core facility with expert staff (see #1 above).
7. Document all steps, reagents, equipment and data analysis methods used in the experiment. Assure that the both the documentation and the data itself are properly stored in a safe data management repository.
8. Acknowledge the Cancer Center Support Grant (P30 CA016086) (if applicable), other grants that support the core, the core (by name), and core staff in publications
Guide to Rigor and Reproducibility for the Pathology Services Core
PSC administration strongly recommends that PIs who would like to run immunohistochemistry (IHC) projects with Digital Pathology tools for epidemiology studies, research projects or clinical trials contact us so that we can discuss the project details as well as important factors that can affect the reproducibility and validity of IHC.
- If available, we recommend using highly validated antibodies that are approved for In Vitro Diagnostic Use (IVD).
- If such antibodies are not available, we recommend using antibodies from vendors who provide information about antibody specificity based on the current requirements (western blot, cell lines, blocking peptides, knockout cells).
- We also recommend using antibodies that have been validated by PSC on previous studies (PSC keeps complete information of all antibodies used in the Lab such as host, clonality, vendor, catalog#, lot #, concentration, storage conditions, and use date).
- For large IHC projects continuing over several years and requiring the staining of study slides in batches, we recommend using only monoclonal antibodies, as such antibodies have less lot-to-lot variability. To reduce the variability of IHC stains caused by different lots we recommend, if possible, using the same lot of the antibody for the entire project.
- We do not recommend using polyclonal antibodies for large epidemiology projects even if the antibody is well validated, because polyclonal antibodies have high lot-to-lot variability and very often they have been discontinued.
- For commercial or homemade antibodies that are not approved for IHC, we recommend testing specificity on a knockout (KO) and a wild-type (WT) cell line. If KO cells are not available, the cell lines with high and low expression of the target or known positive and negative cell lines are acceptable.
- If you’d like to find antibodies, we recommend visiting the following webpages: Human Protein Atlas and Antibody registry.
- PSC offers standardized, fully automated IHC stains to reduce run-to-run variability and operator-related error and to ensure assay reproducibility. All staining units are well-maintained by PSC staff and regularly checked by the vendor based on the annual service agreement. All staining protocols and run events are saved for each slide and are easily retrievable.
- We use highly validated detection kits and high-quality ancillary reagents from leading vendors specializing in IHC products.
- We recommend using the same positive control tissues for an entire project or a mini-TMA containing positive and negative tissues and cell lines. PSC constructed some control TMAs as well as project-specific ones.
- To reduce run-to-run variability of IHC stains caused by different lots of the detection kits or ancillary reagents, we recommend, if possible, staining all study slides with reagents from the same lot. If this is not feasible, before staining the study slides, we perform a QC test of the new kits and reagents on the control tissues/TMA to make sure that there is no unacceptable variation of the staining intensities as compared to the first staining batch. The objective comparison of staining intensities is done by digital pathology tools.
- We always include positive and negative control tissues for each staining run. For large epidemiology projects, when more than 500 slides are stained, we include a positive control for each staining run but a negative control only with the new lot of detection kit/reagents.
- Negative controls run by PSC do not include a primary antibody, which allows us to check any background staining associated (related) with the secondary antibodies and enzymatic reactions.
- We recommend isotype negative controls for the homemade antibodies or antibodies that are not approved for IHC application. Use of isotype controls from the same vendor as the primary antibodies is recommended.
- Staining protocols in PSC are optimized on freshly sectioned or properly stored (nitrogen chamber) control tissues. It is well-known that the antigenicity of some epitopes can be reduced after sectioning if sections are stored at room temperature. For clinical trials or other projects in which tissue sections are received from different institutions and information about sectioning date and storage conditions is not available, we request that the PI provide extra sections for two or three cases to QC the staining protocol on these slides, and only then proceed to stain the study slides.
- All digital pathology platforms in PSC are serviced by the vendor on a regular basis according to the service contract agreement. All imaging platforms are placed on vibration-resistant tables/plates and maintained and calibrated by PSC staff according to the vendor’s recommendations; light bulb or liquid light guide replacement dates are recorded accordingly.
- To ensure the rigor and reproducibility of fluorescent image acquisition between imaging sessions, PSC uses the following protocols:
- Bright-field or multi-channel fluorescent imaging is done on the same platform for all slides belonging to the project if they require automated image analysis.
- First, optimal exposure times are determined for each channel on the control tissue/TMA and on a few study slides, and then the approved image acquisition protocol can be applied to the study slides from all staining batches.
- If a light bulb or liquid light guide has been changed in the middle of a large study, the image acquisition protocol is revised to obtain similar intensities on the control tissues to those obtained before. Alternatively, analysis data is normalized for each staining batch by the intensity of the corresponding control.
- We recommend detection of highly expressed targets in the FITC and Cy3 channels to keep exposure times under 80 milliseconds. This helps to increase the signal-to-noise ratio caused by tissue autofluorescence.
- If two targets are expressed in the same subcellular compartments, we recommend using fluorescent dyes that do not have overlapping emission spectra with our filters (FITC and Cy5).
- Images of the control tissues/TMAs included in each staining batch are used to calculate the coefficient of variation (CV) to determine run-to-run variability of the IHC stains.
- Image analysis and tissue pattern recognition algorithms/macros developed by PSC for images from a specific platform are not translatable to other image formats.