Undergraduates Phenotyping Arabidopsis Knockouts in a Course-Based Undergraduate Research Experience: Exploring Plant Fitness and Vigor Using Quantitative Phenotyping Methods

The CURE module described here is framed in the context of genetic variation for phenotypic plasticity (variation among genetic lines in response to environmental variation, also known as genotype-by-environment effects), and human-influenced environmental change (e.g., increased temperature associated with global climate change). By examining how environmental treatments affect natural accessions and genetic mutants of the plant Arabidopsis thaliana, students have the opportunity to observe that different genotypes of the same species may respond differently to environmental conditions. Further, students observe that mutations can have environment-specific positive, negative, or null effects on complex plant quantitative traits and corresponding phenotypes, such as fitness. Students gain hands-on experience with phenotyping methods (i.e., measuring quantitative traits), experimental design, data management, statistical methods, graphical visualizations, literature searches, working with publicly available databases for model organisms (e.g., TAIR, http://arabidopsis.org, and T-DNA Express, http://signal.salk.edu/cgi-bin/tdnaexpress), and communicating research effectively to peers through writing and oral presentations (Fig. 1; Appendices 1 and 9).

Prior to the course, instructors prepare the experimental design and obtain seeds from the unPAK stock center for mutant, parental (Col-0), and natural lines. (See Appendix 1 for unPAK contact information and for additional instructor support information, including preCURE set up, control line use, suggestions on treatment implementation, tips on randomized complete block design, guidance on building experimental designs aligned with pedagogical and research goals, and managing space constraints; see also Fig. 1, weeks −2 and −1). Instructors vernalize (cold-treat) seeds on wet filter paper for a week (Fig. 1) and then transfer seeds to individual pots (see Appendix 1 for materials). Instructors grow plants in pots arranged in growth chambers in a randomized block design. After three to four weeks of growth and in the second week of the semester, each pair of students measures six replicates per treatment of two lines each, either 1) COL 70000 and a T-DNA mutant line or 2) a pair of natural accessions. Option 1 enables students to investigate reproductive effects of a disrupted gene (T-DNA mutant line) and environment-specific mutant effects. Option 2 enables students to compare two naturally-occurring Arabidopsis lines to view how natural genetic variation affects plant responses to environmental variation. All of the class’s plants constitute a single temporal block of a stratified random experimental design as a contribution to unPAK research efforts for specific traits, and students explore additional phenotypes based on their interests. Thus, multiple sections of a course across semesters contribute to replication of an experiment. Inclusion of the COL70000 wild-type and natural accessions allows for comparison of data collected across CUREs implemented within the unPAK network (Fig. 1, Appendices 2 and 9).

During the first week of the module, students read background material to develop hypotheses of environment responses for their lines (Fig. 1; Appendix 2). During the second week, students make initial measurements of quantitative plant phenotypes, specifically rosette diameter (Appendices 1 and 3), which follow standard unPAK protocols and can later contribute to the curated central unPAK database (see http://arabidopsisunpak.org for details). Between measurements, while plants are growing, students expand literature searches and refine their hypotheses as they develop expertise (Fig. 1; Appendix 6). Instructors may engage students in activities unrelated to unPAK, with other learning goals, during this period. Also during this period, instructors water and tend to plants; however, this could easily be done by students at institutions where students have regular access to growth facilities. Students measure fruit production in week seven. Because the students observe plants both in early growth stages (class week 2; Appendices 2 and 3) and late growth stages (class week 7; Appendices 4 and 5), they have the opportunity to observe how plant traits at distinct life stages are affected by the environmental treatments and whether the two genotypes respond differently to the treatments. In week nine, students graph and analyze their data (Appendix 6), using these results to produce a primary literature style paper on their findings (Fig. 1; Appendices 7 and 9). Throughout the module, instructors emphasize the use of the data to advance scientific knowledge within the Arabidopsis community and describe how careful measurements and well-curated data can help to further that knowledge.

As mentioned above, students are required to measure two unPAK-defined phenotypes, rosette diameter and fruit production, following specific protocols for measurement (Appendix 1). Through this process, students gain experience with formal protocols useful for addressing questions of genotype by environment. Students’ datasets receive two peer reviews, an instructor review, and an unPAK database manager review prior to inclusion into the central unPAK database (arabidopsisunpak.org). To validate student data and check for consistency and quality, datasets are a) checked for the direction of the overall sample response to treatment (e.g., if plants were observed to be larger in one environment, we would expect the general student dataset to reflect that direction), b) screened for extreme outliers indicative of typing errors, c) examined for order of magnitude differences in trait values for a particular group indicative of errors in measurement units (e.g., cm versus mm), and d) checked for matches between the plant IDs students submit and the plant IDs assigned to them. When multiple sections of students measure the same plants, as often happens in unPAK courses with multiple lab sections, measurements are evaluated for consistency. If students fail to follow measurement protocols, do not properly submit data and metadata in the unPAK template (Appendix 9), have incomplete columns or metadata missing, or include units in the column with quantitative values instead of recording them in the meta-data, the data fail quality control steps and are not included in the database. This process contributes to quality control in data submission (Appendix 9) and confirms that appropriate control lines were included in each experiment. This is essential to ensure the quality and utility of the data once it is in the database where it will be used by other unPAK researchers for further analyses or to inform future research. Successful completion of the project and communication of findings indicates successful accomplishment of learning objectives 1 and 2. Individual student assessment is based on a final presentation of results in a literature-style paper (Appendices 6 and 8).

Students’ successful collection of plant quantitative phenotypes, record-keeping, graphing, statistical analyses, and submission of data to the unPAK database constitute evidence of this objective. All students collect data on rosette diameter and fruit number after being trained in the use of calipers and in plant growth patterns particular to this species. As formative feedback on students’ measurement efficacy, peer partners check an early subset of each other’s measurements by observing their partner, commenting on technique, and suggesting any necessary improvements or seeking additional instructor guidance. Students’ data are then subject to the rigorous quality control procedures described above. Given the level of scrutiny each sample is subject to, successful completion of the above steps and submission of complete data lines to the database can be taken as evidence of success in learning how to measure quantitative plant phenotypes and communicate findings.

Students are evaluated as to whether they have accomplished this objective using their final writing assignment, a scientific-style paper detailing their rationale, methods, and results. In the paper introduction, students must discuss prior research linking their mutants or natural populations (genotypes) to traits characterized in the literature (phenotypes). Then, in the results section, students must report on their findings, describing how their chosen mutants or natural populations (genotypes) responded (phenotypic responses) to variation in environments in their text and use graphical and statistical procedures. Finally, students are required to interpret and explain their results, including discussing the links between genotype, phenotype, and environment in the discussion (see Appendix 7). Students are then graded on their ability to make these connections (see Appendix 7 rubric), indicating their successful ability to explain the link between genotype and phenotype and responses across environments.

The opportunity to make a scientific discovery is an important design feature of CUREs for three reasons: 1) students have the potential to actually make a discovery that might contribute to science, 2) pursuing a discovery that is new and relevant to the scientific community may motivate students to engage, since the prospect of discovery is exciting, and 3) pursuing an unknown answer to a scientific question, regardless of whether students support their hypothesis or make a new “discovery,” necessitates student engagement with the uncertainty inherent to science. Students must be able to logically explain their own results; they cannot look to past research or experiments to find the correct answer as they may be able to do in more traditional labs. This leads to learning regardless of the results of the experiment. The discovery scale in the Laboratory Course Assessment Survey (LCAS) measures students’ opportunities to make relevant discoveries (17). This scale consists of five items that ask students to report on whether they had opportunities to make new discoveries or develop new ideas of relevance to the scientific community. Response options for each question range from 1 (strongly disagree) to 6 (strongly agree). This scale was given to students at the end of the unPAK CURE, and the course average was compared with scores collected from a national sample of CUREs using t-tests in order to assess whether students reported opportunities to make scientifically relevant discoveries and whether these opportunities were comparable with those offered by other CUREs.

Development of students’ science self-efficacy is an important outcome as it is correlated with students’ long-term persistence in STEM (22, 23). Chemers’s science self-efficacy scale consists of six items that ask students to rate their confidence in their ability to do scientific tasks on a scale of 1 (not at all confident) to 5 (absolutely confident) (24). We used a pre-/post-test approach to measure changes in science self-efficacy over the duration of the course. We assessed differences in students’ scores for self-efficacy using a matched-pairs t-test performed on pre- and post-test data.

To assess LO3 and LO4, we followed appropriate protocols for informed consent and confidentiality in data collection (IRB, College of Charleston # IRB-2014-08-29-154322).

Variation in plant phenotypes was uncovered in data collected by unPAK students in the classroom. Students contributed this data to a national database used by researchers across the unPAK network, demonstrating students’ ability to measure quantitative plant phenotypes and communicate their findings. Out of 123 pairs of student data reviewed from a random selection of sections between 2012 and 2016 at CofC, 80.5% of student pairs’ data files met data inclusion standards and were subsequently included in the unPAK database. This assessment underscores students’ success in achieving gains in their ability to execute quantitative phenotypic methods. Please see the sample plant data section for an example of student results that met data inclusion standards and indicate achieving LO1.

Students successfully demonstrated an ability to explain the link between genotype and phenotype and variation across genotypes in phenotypic responses across environments. The final paper (see Appendix 7) is the most accurate and authentic measure of this learning outcome since students are required to construct arguments that link genotype, phenotype, and environment in the paper (see Suggestions for determining student learning section, above, for details). Any student failing to execute paper tasks linking genotype, phenotype, and environment is unlikely to receive above a maximum score of 66% on their final project (see the rubric and point structure in Appendix 7). Since 66% is considered a barely passing grade, passing the final project with a 70% or better indicates accomplishment of LO2. Of students who completed the assignment and the course (over 16 sections each with ~20 students per section), 97.7% received an acceptable passing grade on the final paper, indicating achievement of this outcome. Among the paper components most closely related to LO2, students most commonly struggled with the explanation of their results in the context of prior work, indicating that future work needs to focus on class examination of results and discussion sections of published literature. Future classes could incorporate more opportunities to practice this skill. For example, students might read additional relevant literature and engage in a class discussion about how the literature is important to their work, with specific coaching on how they might reference the literature in their final report. The module written assignment and activities represent 16% of the total points for the four-credit course at CofC.

Students had opportunities to discover something new to the scientific community. The course provided students with opportunities to make relevant scientific discoveries comparable with students in a national sample of CUREs. The discovery and relevance scale used to measure students’ opportunities to make discoveries consisted of six questions with five possible responses each (summed score range: 5 to 30, Cronbach’s alpha = 0.87). Opportunities for relevant discovery reported by unPAK CURE students (mean = 25.14, sd = 5.16) did not differ significantly from students in the national sample of CUREs reported in previously published work by Corwin and colleagues (9, 17) (n = 72, mean = 24.35, sd = 4.04, p = 0.408). In addition, students showed significantly more opportunities for relevant discovery than students in a national sample of traditional labs (n = 60, mean = 20.77, sd = 5.82, p < 0.001). These results demonstrate that the unPAK CURE is similar to other CUREs nationally in offering opportunities for relevant discovery, which is predicted to increase students’ motivation and contribute to their persistence in science and, importantly, offers students the opportunity to work on a problem with an unknown answer. This constitutes a learning opportunity regardless of whether students’ results support a specific hypothesis (17). However, it is important to note that, overall since 2012, over 150 mutant lines have been screened in unPAK CUREs, generating new information that constitutes unPAK’s novel discoveries.

Students gained confidence in their ability to do scientific research. The self-efficacy scale used to measure students’ changes in self-efficacy consisted of five questions with six possible responses each (summed score range: 6 to 30, Cronbach’s alpha = 0.86, 26) given to course sections of 20 across two semesters. Students made significant gains in science self-efficacy, with a pre-course mean score of 21.99 (sd = 4.71) and a post-course mean score of 24.20 (sd = 5.40; p < 0.001). Such an outcome is a common result of one-on-one mentored research participation and has been characterized as an indirect contributor to persistence in science, since it precedes outcomes that directly predict persistence (22, 23, 27).

After implementing the CURE, instructors (n = 6) were queried about their experience (28; further details in Appendix 11). We asked, “Why did you choose to implement a CURE? What goals did this module accomplish that a non–research based course might not have accomplished? What was the most important outcome of this module for your course?” All faculty agreed that the activities were valuable for their students and were well designed while offering flexibility. All stated they would use the CURE again in their course (Supplemental Table, Appendix 11).