Ensure you have a good mandate and a supportive institution to take on the question.
Use the fewest assumptions possible to design your experiment. Assume that half or more of the existing literature (and tools) may not be reproducible or may have shortcomings.
Design experiments that give the maximum informational value. Testing whether just one treatment cures a disease yields only one bit of information: yes or no. If you can test which of five treatments has efficacy, that is a much more informative experiment.
Use orthogonal approaches to ask the same question. This avoids the artifacts and biases innate to each approach and allows for orthogonal validation.
Well-trained scientists and their passion are important rate-limiting assets of our society. Given how many important scientific questions there are to ask, reducing failure and ensuring each of you succeeds is important even beyond your personal career.
There are a lot of reasons why a research project can fail, and some of them are avoidable. For biological research that focuses on genes, assessing ‘which cell’ and ‘how essential’ is critical. A brilliant hypothesis about a gene’s function can crumble if the gene isn’t expressed in the right tissue or if it’s so vital that any perturbation is lethal. Fortunately, a wealth of publicly available data can help you de-risk your project from the outset. This post will walk you through a powerful trifecta of online resources—GTEx, Allen Brain Atlas, and gnomAD—to constrain your research scope and build a more robust hypothesis.
Step 1: Where in the Body is Your Gene of Interest? Ask the GTEx Portal
Before you even think about complex experiments, a fundamental question needs an answer: Is your gene of interest expressed in the tissue you plan to study? The Genotype-Tissue Expression (GTEx) portal is an invaluable resource for exploring gene expression across a wide array of human tissues.
The GTEx project has collected data from a multitude of post-mortem donors, providing a comprehensive look at the normal expression landscape. You can visualize both bulk RNA-sequencing data, which gives an average expression level for a tissue, and single-cell RNA-sequencing data for a more granular view.
Search for Your Gene: Use the search bar at the top to enter your gene of interest. You can use a gene symbol, Ensembl ID, or other identifiers.
Analyze the Expression Profile: The results page will display a variety of plots. The most immediately useful is often the “Gene Expression” violin plot. This shows the distribution of expression levels for your gene across all available tissues. Look for high median expression in your tissue of interest. The spread of the “violin” indicates the variability in expression across different individuals.
Explore Splicing and eQTL Data: For a deeper dive, look for splicing patterns that would change gene function under ‘Exon expression’, and for eQTLs which can reveal how some humans have significantly different expression coupled with SNPs.
By starting with GTEx, you can quickly confirm if your gene is present in the right biological context, preventing you from pursuing a project in a system where your gene is silent.
Step 2: Zooming into the Brain with the Allen Brain Atlas
If your research focuses on neuroscience, the Allen Brain Atlas is an unparalleled resource. It provides a detailed, genome-wide map of gene expression in both the mouse and human brain. This allows you to investigate your gene’s expression with high spatial resolution.
Navigating the Allen Brain Atlas for Your Research:
Choose Your Species: You can explore data from both mouse and human brains. The mouse data is particularly comprehensive.
Search for Your Gene: Similar to GTEx, you can search for your gene of interest.
Interpret the Expression Patterns: The atlas provides stunning single cell RNA-seq (scRNA-seq) datasets and in situ hybridization (ISH) images that both visualize where your gene’s mRNA is located within the brain. You can see if it’s enriched in specific regions, like the hippocampus or cerebellum, or in particular cell types like PV+ basket neurons. The platform also offers tools for differential searches to find genes enriched in one brain region compared to another.
The Allen Brain Atlas helps you refine your hypothesis by pinpointing the specific neural circuits and cell populations where your gene might be active, adding a crucial layer of spatial detail to your research question.
Step 3: Gauging Gene Essentiality with gnomAD pLI Scores
Now that you know where your gene is expressed, you need to understand how important it is for basic biological function. The Genome Aggregation Database (gnomAD) is a massive collection of exome and genome sequencing data from a diverse range of individuals. A key metric derived from this data is the probability of being Loss-of-Function Intolerant (pLI) score.
The pLI score estimates how tolerant a gene is to mutations that would likely abolish its function. The score ranges from 0 to 1:
pLI score close to 1 (typically ≥ 0.9): This indicates that the gene is highly intolerant to loss-of-function mutations. Such mutations are rarely seen in the general population, suggesting that they are likely to cause a severe, often embryonic, phenotype. These genes are considered highly constrained.
pLI score close to 0: This suggests the gene is tolerant to loss-of-function mutations, and inactivating one copy of the gene is likely not detrimental.
Search for Your Gene: Enter your gene’s name in the search bar.
Find the pLI Score: On the gene’s page, you’ll find a “Constraint” table. This table includes the pLI score.
Interpret the Implications:
High pLI (≥ 0.9): If your gene has a high pLI score, it’s a double-edged sword. On one hand, it suggests the gene is functionally important, which means that gene manipulation will easily lead to observable phenotypes. On the other hand, it may be difficult to study using traditional knockout or knockdown approaches in model organisms, as this could be lethal. You might need to consider more subtle genetic manipulations, such as studying hypomorphic alleles or using inducible systems.
Low pLI: A low pLI score might suggest that a complete loss of function is viable, making it a more tractable target for knockout studies. However, it also suggests the gene is not essential for an organism to develop, mate, and reproduce. If your research aims to find genes that must contribute to these processes, it will likely be fruitful to look for other genes that have a higher pLI value. This doesn’t mean the gene is unimportant; it could have more subtle roles, be part of a redundant pathway, or be critical only under specific conditions that are unlikely to provide an evolutionary advantage, like neurodegeneration.
While gnomAD has more recently introduced the observed/expected upper bound fraction (LOEUF) score as a more continuous measure of constraint, the pLI score remains a widely used and intuitive metric for a first-pass assessment of gene essentiality.
Synthesizing the Data for a Stronger Hypothesis
By integrating information from these three powerful resources, you can build a much more robust and defensible research plan.
Imagine you have a hypothesis that Gene X is involved in memory formation in a specific type of neuron in the hippocampus.
GTEx and Allen Brain Atlas: You first check GTEx and find that Gene X is indeed highly expressed in the brain, and single-cell data confirms its presence in the neuron type and layer your hypothesis implicates.
gnomAD: Since memory is essential for an organisms’ survival and reproduction, we would expect a high pLI value for any gene critically involved in learning and memory. You look up Gene X in gnomAD and find it has a pLI score of 0.95. This tells you the gene is likely essential. Therefore, a full knockout mouse might not be viable. This insight guides you to design a more sophisticated experiment, perhaps using a conditional knockout strategy to delete the gene only in the hippocampus at a specific time point, or using CRISPRi to partially knockdown its expression.
By taking these preparatory steps, you’ve significantly de-risked your project. You’ve confirmed your gene is in the right place at the right time and have anticipated potential challenges related to its essentiality. This proactive approach allows you to refine your hypothesis and design more targeted and meaningful experiments, ultimately steering your research away from avoidable dead ends and towards success.
