The emergence of large-scale data collection in both molecular biology and biodiversity sciences is increasing the demand for scientists proficient in computational data analysis (1-3). The response to this demand are lessons or entire classes that teach bioinformatics to students with varying familiarity with programming, statistics, and even basic computer skills. Instructors in bioinformatics classes face the daunting task of making high-level, abstract computational concepts accessible to students who have no frame of reference for basic concepts in computer science, and who may even be acutely intimidated by them.
An obvious solution to encourage student learning of complicated concepts is to relate these foreign terms and processes to things with which students are familiar. Such metaphors and analogies have been widely used throughout the history of biology to elucidate concepts, and their use in science classrooms has been explored and evaluated extensively (e.g., 4). Although "metaphor" and "analogy" are often used interchangeably in science education literature, analogies are generally more structured and require explicit connections between concepts (5).
Analogies are often incorporated into both textbooks and classroom teaching to explain myriad biological topics, including cell/molecular biology, genetics, and evolution (e.g., 6-8). Moreover, analogies can act as an introduction to scientific modeling (a formal method of scientific inquiry that replicates the natural world in a conceptual framework), as modeling requires students to draw on reasoning by analogy (9). Analogies are also frequently used in popular media to describe complex topics related to bioinformatics (10). Despite this prevalence, analogies have a contentious history in biology education (4). For example, use of analogies may sometimes confound student understanding of a topic (11), both by reinforcing previous misconceptions and failing to build an appropriate framework for abstract concepts.
Despite possible drawbacks, I contend that analogies are an invaluable tool to explore and improve students' understanding of foundational concepts in bioinformatics. An appropriately designed analogy should bridge the gap between a familiar, tangible concept and a new idea that may be unfamiliar and/or intangible. The following example describes the application of an analogy to introduce students to an essential aspect of computational thinking.
By asking students to relate complex or abstract concepts in biology and computer science to common, familiar events and processes, analogies promote active engagement in learning. Describing and expanding an analogy require students (individually or in pairs) to evaluate both their own capacity to understand the analogy, as well as their classmates' attempts to describe bioinformatics processes. Finally, breaking down the analogy is an essential step for students to personally identify potential pitfalls or misconceptions about the process being described.
Analogies can be used in various ways in science classrooms. The in-class "making toast" example is useful as an explorative exercise, and instructors can quickly and easily gather informal feedback about the extent to which students can conceptualize procedures and steps in algorithms. However, the concept of analogies can also be incorporated into formal assessment. Following the in-class introduction to computational thinking, I asked students to describe their own common, daily task that can represent a bioinformatics workflow (see instructions and associated rubric in S1). At the end of the semester, I included a question in the summative (cumulative) assessment for the class that asked them if they could identify any additional ways in which their selected analogy breaks down. Because the initial "making toast" exercise occurs at the beginning of the semester, students are much better equipped in this summative assessment to describe specific relationships, extensions, and drawbacks between their analogy and bioinformatics workflows.
The use of analogies encourages students to share knowledge about diverse cultures and experiences in the science classroom. In the "making toast" example, students are reminded about plurality in food choices, both in terms of diet (what foods are eaten with toast, including dietary restrictions) as well as preparation (what equipment is available for making the toast). Assessing student learning through formation of their own analogies encourages students to share experiences important to them. For example, students assigned to analogize a daily task to a bioinformatics workflow chose processes ranging from quilting to baling hay. Finally, some genomics analogies have been extended to physical models, allowing visually impaired and blind students to engage in lessons about gene sequencing (12), and other analogies could be similarly developed to encourage inclusivity.
A challenging aspect of learning bioinformatics is developing requisite skills in computational thinking, a type of analytical thinking requiring abstraction at multiple levels of complexity (13). Given that analogies require abstraction of the familiar to relate the unfamiliar, this is an example well-suited to the strength of analogy use in bioinformatics education. This exercise introduces students to computational thinking through the process of making toast. This process of food preparation is commonly applied in computer science to help students learn about algorithms, or the step-by-step process of solving a problem with a computer (14-15). In this example, the process of making toast is generalized to represent steps in the process of a bioinformatics/data analysis workflow. The steps described below focus on the second phase of the Focus, Action and Reflection (FAR) model of teaching and learning science with analogies (16). This teaching process is designed for biology students at the high school, undergraduate, or graduate level, and is structured to highlight particular challenges to student understanding of bioinformatic workflows.
Introducing the analogy
The introduction of the analogy should help students actively practice skills in computational thinking, as well as develop an understanding for how parts of a bioinformatics process work together. Show students a picture of a loaf of bagged bread and a toaster (Figure 1) and ask them to think about how they would describe the process of making toast. Have students work with a partner to write down the steps required. Some students may have questions, such as "Is the toaster plugged in?" and "How much are we assuming?" I generally gesture toward the picture and tell them to base their steps on what they see there. You may choose to give them additional parameters, ranging from very specific ("Imagine a child in your house needs instructions") to very general ("Imagine you are telling a Martian/extraterrestrial how to make toast").
After students are satisfied with their lists of steps, help students summarize their results as a class. This is an appropriate time to begin explaining the connections between the analogy and the process of bioinformatics (a workflow or algorithm). Most students will quickly identify that the steps they outlined are like lines of computer code. You may choose to make connections more explicit by describing untoasted bread as the raw data, and the final toast as the output from the analysis. Questions for further discussion may include the following:
How many steps did you outline? Most students will identify 5-7 steps; some students will specify a dozen or more. The steps generally differ in the level of detail included in the process. You can relate this to the specificity of a bioinformatics analysis, and point out the assumptions students made when describing their steps.
What assumptions did you make about the start and end points for the process? Most students will start with a step to untwist the fastener on the bag of bread, but some will include steps for walking to the place on the counter where the bread is located. Similarly, some students will leave the bread in the toaster, while others will include steps to eat the toast. Bioinformatic analysis requires specific input and output, identified from the start of the experiment.
Did anyone include a way to test whether the bread is toasted to the appropriate level? Some groups may have specified the length of time for the bread to stay in the toaster. A few students may use color of the toast to identify whether it is completely done. Validation is important in bioinformatics, so assessing the success of the process is essential.
Did anyone include conditional statements to make toast? Some students may have used if/then steps to determine the process for making toast. These are directly relatable to common coding methods (for loops, if/then, if/else) in algorithms and bioinformatics analysis, and are essential to offering flexibility and transferability to code.
Expanding the analogy
The expansion of the analogy will allow you to more explicitly relate parts of making toast to a bioinformatics process by offering alternative conditions for the steps students have described. Ask pairs of students to trade their list of steps with other student pairs. Ask students to compare their methods with the other group to gain better understanding of the variation in how steps are described. Do the other group's steps accomplish the process effectively? You can then select from a series of different options to modify the process of making toast and deepen students' understanding:
Type of bread (Figure 2A). Ask how many more steps are required to complete the process of making toast if the initial materials differ. Some steps, like cutting the loaf of bread into slices, may seem trivial, while others, like baking bread, may require another whole series of steps. Relate this to steps that may be required to pre-process (filter, parse, or quality control) data prior to analysis.
Toasting apparatus (Figure 2B). Ask students what needs to change in their steps to allow for different methods of toasting the bread. A toy toaster oven, for example, may not have enough power to bake bread to completion, while an open-fire oven may need to be managed more carefully. Relate this to equipment available for bioinformatics analysis, and contrast a desktop/laptop computer with a high-performance computing resource.
Resulting food (Figure 2C). Ask students which steps need to be changed, or if additional steps need to be added, if the desired food is modified from basic toast. The case of toaster strudel is akin to starting with a different type of data, while adding an egg to the toast is similar to combining data types.
What happens when something goes wrong? Ask students what steps are required if there is an interruption or deviation in the process. For example, what if the electricity goes out? Can the process be restarted? These questions can help students prepare themselves for troubleshooting their own bioinformatics workflows later in the semester.
Breakdown of the analogy
Understanding how the analogy inaccurately represents the biological or computational process is essential to correcting possible misconceptions in student learning. Remind students that all analogies fail in at least one way. Ask them to describe how making toast fails to accurately represent a bioinformatics process. There are many possible student responses, depending on their familiarity with data analysis and computer science, and many of them will likely be superficial (i.e., "You can't eat the results of a bioinformatic analysis"). The following are example prompts that can help students think more critically about the suitability of the analogy.
Reproducibility. If the final result from toasting bread isn't acceptable (i.e., not to our tastes), we have very limited ability to improve the outcome. While we can continue to toast the bread if it isn't crispy enough for our liking, we cannot untoast the final product. At best, we must start with a completely different piece of bread. When performing a bioinformatic analysis, however, it is possible (and even necessary, in many cases) to begin the workflow over again from the exact same data. Therefore, bioinformatics is reproducible in a way that making toast is not.
Testability. When making toast, we have clear and precise expectations about the desired outcome, and we choose the time and nature of toasting to meet those desires. Bioinformatic analysis, on the other hand, is generally associated with testing a specific hypothesis to answer a research question. While we can choose the methods with which use to obtain the answer, we cannot choose the final answer. Whereas bioinformatic analysis allows testing of specific questions, making toast is specific to a final outcome.
There are opportunities to incorporate analogies into bioinformatics curriculum from both biological and computer science perspectives (Table 1). The dominance of genomics as a commonplace approach in biological science is a prime example for the application of analogies in teaching, as biology teachers are increasingly compelled to expose students to topics in the field. However, genomics includes processes that are both complex and emergent (patterns at the organismal level arise through interactions among smaller units, like genes), which can be difficult to convey to students. Analogies have been effectively used in other fields to convey content with similarly difficult concepts (17-18), and analogies are especially useful for helping students make sense of abstract concept in molecular biology (11).
Whereas metaphors related to the molecular biology and the genome are related to biological understanding, analogies can also be used to describe conceptual and procedural methodologies commonly employed in computer science (Table 1). Analogies, in fact, can help bridge the gap between biology and computer science by highlighting the commonalities in their abstract concepts. One obvious relationship between these two disciplines is mathematical modeling. The development, testing, and application of models is a cornerstone of computational biology. By definition, analogies are essentially an unscientific but relatable model for understanding a concept. When employed appropriately, analogies can serve as a model for teaching scientific models to students (19).
The example presented in this article represents one possibility for a thorough, structured approach to teaching biology students about the process of performing bioinformatics analysis using analogy. Analogies can be applied to myriad topics in many different ways to educate students about the connections between biology and computer science (Table 1). Peer discussion, collaborative development of self-generated analogies, and reflection yield the best return on student understanding from analogies (20-21). Thus, the most effective use of analogies as a teaching strategy will intentionally incorporate those activities. Extended analogies offer a conduit for students to develop deeper understanding of abstract computational and biological content, as well as scientific reasoning skills.
S1.Making Toast-Assignment and assessment rubric for student creation of their own analogy
I thank Mike Barker for introducing me to the "making toast" exercise as I planned my first bioinformatics lessons, and Matt Dorrance for helpful discussion in developing the analogy further. Finally, many thanks to my undergraduate and graduate bioinformatics classes in spring and fall 2015 for helping me refine implementation of this strategy.
- Tan TW, Lim SJ, Khan AJ, Ranganathan S. 2009. A proposed minimum skill set for university graduates to meet the informatics needs and challenges of the "-omics" era. BMC Genomics. 10 Suppl 3:S36.
- Greene AC, Giffin KA, Greene CS, Moore JH. 2015. Adapting bioinformatics curricula for big data. Brief Bioinform. 1-8.
- Smith JT, Harris JC, Lopez OJ, Valverde L, Borchert GM. 2015. "On the job" learning: A bioinformatics course incorporating undergraduates in actual research projects and manuscript submissions. Biochem Mol Biol Educ. 43:w154-161.
- Venville GJ, Treagust DF. 1997. Analogies in biology education: A contentious issue. Am Biol Teach. 59:282-287.
- Niebert K, Marsch S, Treagust DF. 2012. Understanding needs embodiment: A theory-guided reanalysis of the role of metaphors and analogies in understanding science. Sci Educ. 96:849-877.
- Bean TW, Searles D, Singer H, Cowen S. 1990. Learning concepts from biology text through pictorial analogies and an analogical study guide. J Educ Res 83:233-237.
- Dagher ZR. 1995. Review of studies on the effectiveness of instructional analogies in science education. Sci Educ. 79:295-312.
- Kirkpatrick G, Orvis K, Pittendrigh B. 2002. A teaching model for biotechnology and genomics education. J Biol Educ. 37:31-35.
- Svoboda J, Passmore C. 2013. The strategies of modeling in biology education. Sci & Educ. 22:119-142.
- Condit CM. 1999. How the public understands genetics: non-deterministic and non-discriminatory interpretations of the "blueprint" metaphor. Public Underst of Sci. 8:169-180.
- Tibell LA, Rundgren CJ. 2010. Educational challenges of molecular life science: Characteristics and implications for education and research. CBE Life Sci Educ. 9:25-33
- Butler CR, Bellow J, York A, Orvis K, Pittendrigh BR. 2013. Genomics analogy model for educators (GAME): Fuzzy DNA model to enable the learning of gene sequencing for visually impaired and blind students. Sci Educ Rev. 7:1-9.
- Wing JM. 2008. Computational thinking and thinking about computing. Philos Trans A Math Phys Eng Sci. 366:3717-3725.
- Aspinall B. 2013. Now that's how you make toast! Using computer algorithms. Brian Aspinall, CV. http://brianaspinall.com. Accessed February 1, 2016.
- Wikiversity, The School of Computer Science. 2015. "Introduction to algorithms/What is an algorithm. https://www.wikiversity.org. Accessed February 1, 2016.
- Treagust DF. 1993. The evolution of an approach for using analogies in teaching and learning science. Res Sci Educ. 23:293-301.
- Slotta JD, Chi MT. 2006. Helping students understand challenging topics in science through ontology training. Cogn Instr. 24:261-289.
- Jacobson MJ. 2001. Problem solving, cognition, and complex systems: Differences between experts and novices. Complexity. 6:41-49.
- Harrison AG, Treagust DF. 2000. A typology of school science models. International Journal of Science Education. 22:1011-1026.
- Coll RK, France B, Taylor I. 2005. The role of models/and analogies in science education: implications from research. Int J Sci Ed. 27:183-198.
- Haglund J. 2013. Collaborative and self-generated analogies in science education. Stud Sci Educ. 49:35-68.