CUREing biochemistry lab monotony

Abstract

Undergraduate research experience is critical to success in post-graduate research settings. The recent movement away from “cookbook” style labs to course-based undergraduate research experiences (CUREs) in undergraduate laboratories has allowed universities to provide inclusive research experience while bypassing the limitations of extracurricular apprenticeships. This paper describes an upper-level biochemistry CURE designed to provide students with an introductory experience to graduate-level research by studying a suspected DNA helicase. This CURE is designed to span multiple semesters, where each student cohort builds upon the work of previous semesters. Pre- and post-course surveys were employed to assess student confidence in bench skills, perceptions of the course, and project ownership. The results show that the incorporation of lab meeting-style recitations and poster presentations led to higher project ownership, while overcoming troubleshooting was a significant challenge. Furthermore, confidence in every experimental technique increased significantly in all but one instance. Based on these results, this CURE is providing students with a realistic experience in graduate-level research.

Graphical Abstract

Table of Contents Figure. Created in BioRender.com

Introduction

For decades, undergraduate biochemistry labs have been providing an experience in the broad techniques of biochemical research utilizing an expository approach where students are guided to a known outcome¹. While this cookbook-style approach is well-established and easily implemented, it is an unrealistic model of research in a graduate or professional environment^2,3. This shortcoming led the American Association for the Advancement of Science⁴ and President’s Council of Advisors on Science and Technology⁵, among others, to prescribe curriculum updates in undergraduate laboratory methodology to incorporate more research-based instruction.

Traditionally, undergraduate research experiences are gained by conducting an independent project within a faculty member’s research group^6,7. Although such research is an excellent resource for undergraduates, it is often pursued outside of the course load⁶ and is limited by the number of positions available⁷. The effect of these limitations is that many students may not have the opportunity to experience research, the impact of which disproportionately affects students who are underrepresented minorities (URMs)⁸.

In recent years, course-based undergraduate research experience (CURE) courses have been increasingly incorporated to address these shortfalls⁹. Numerous studies have shown that such inquiry-based approaches result in greater internalization of scientific concepts, increased self-confidence in physical skills, data analysis, and interpretation, as well as development of critical thinking skills involved in synthesizing scientific hypotheses and assertions^7,10–13. Completion of research-based courses also has been shown to increase retention of STEM students^11,14, quality of their research experience¹⁵, and access to scientific mentorship¹⁶. These attributes correlate closely with increased self-identification as a scientist by students, especially URMs, and lead to higher likelihood of persistence of students from diverse backgrounds within STEM careers¹⁷.

As an R1 research university, the University of North Carolina at Chapel Hill (UNC) sought to enhance undergraduate research opportunities through incorporation of CUREs in 2017. To implement a CURE for upper-level undergraduates that is representative of academic and industrial research, we created an undergraduate biochemistry laboratory course that provides a research experience comparable to a graduate-level project that is carried out over the course of multiple semesters by different student cohorts. This paper outlines the CURE structure and instructional strategies employed in the course and examines how implementing a CURE model impacts students’ sense of project ownership¹⁸, research skill development, and perceptions of scientific work. Overall, this course had positive impacts on student attitudes toward the skills and projects presented in this course.

Brief introduction to UNC Biochemistry lab

At UNC, the biochemistry lab is a three-credit course primarily taken by third- and fourth-year chemistry majors focused on biochemistry. The class is divided into two sections, each with a maximum of 16 students. Each section meets twice weekly for four-hour periods, as well as a single hour-long recitation period attended by both sections. In addition to a faculty instructor, there are two graduate teaching assistants (TAs) for each section. The traditional aim of the course is to introduce students to the skills involved in the biochemical study of enzymes. For over two decades, this class conducted experiments using previously characterized proteins with known outcomes. Conversations with past students provided anecdotal feedback that suggested students felt they were doing work for the sake of work, which bred disinterest. The goal of this CURE is to create a more realistic, interdisciplinary research environment with activities and concepts from biochemistry, biophysics, and molecular biology. Two main goals of this course are that students come away with realistic expectations of a career in research science and that the research the students perform over multiple semesters will be included in peer-reviewed publications.

Choosing a Research Project

General thoughts on choosing a topic

Using a “backward design” approach^19,20, the research question is designed such that it utilizes the core skills that students need to learn. Key skills for biochemistry students encompass molecular biology, protein purification, and enzyme kinetics. Because the benchwork is being conducted by inexperienced scientists who are restricted in lab time by the course schedule, an essential consideration in this implementation is the feasibility of addressing the research question(s) in an undergraduate lab setting. Experiments that are prohibitively difficult or extend beyond the time constraints of lab periods may not be well-suited for a CURE program. Institutional resources (i.e., availability of instrumentation, funding, and instructors/teaching assistants) also will dictate project feasibility and scope. Furthermore, because research is often unpredictable and the interpretation of the results requires some level of expertise, the instructor should have extensive familiarity with the knowledge base and techniques to assure sufficient course progression. A knowledgeable instructor who can guide students in their data analysis can bolster the students’ confidence in troubleshooting and repeating experiments. Finally, it is important to design a project that has published controls the students can run in tandem in case significant obstacles occur during the research project, ensuring that students learn the core skills even if experiments fail during the semester.

The majority of graduate-level research projects cannot be completed in a single semester; therefore, the new curriculum should be designed to build each semester’s experimental workflow on the results obtained by previous students. This design philosophy results in a dynamic syllabus that attempts to answer those questions that remain incomplete or unaddressed. The adaptability of this curriculum provides the opportunity to reinforce those techniques that the students find particularly challenging and introduce more non-traditional experimental design that might otherwise be precluded by the constraints of traditional course design.

The updated curriculum emphasizes the development of a “biochemist’s toolkit.” Students should understand, in both theory and practice, the array of experimental techniques at their disposal to investigate any problem that they might encounter in a research career. This toolkit includes the core molecular biology skills standard to the isolation of any enzyme – PCR, plasmid cloning, bacterial culturing, separation methods – while also including more advanced experiments that may not be covered by other courses. The concepts covered in this course are summarized in Table S1.

The aim of this approach is to have students work through the development of a research project as they would in a graduate environment: using available resources (e.g., software, literature sources) to understand the current knowledge base, designing experiments, and collecting and analyzing data to test their hypotheses. The experiments performed are a combination of instructor- and student-driven experimental designs (Table S2). Furthermore, the outcomes of their initial experiments influence the methodology of later studies (e.g., reaction temperature or concentration of reactants). While the scope of these experiments is naturally limited to the techniques presented to them in the course, the process is broadly applicable to identifying and solving research problems in a professional research environment.

Approach

Taking the above considerations into account, the instructors selected the research project rather than having students identify their own project. Thermus aquaticus (Taq) UvrD, a homolog of the well-characterized Escherichia coli (E. coli) UvrD DNA helicase that is a key component in the nucleotide excision repair pathway^21–24, was selected as the experimental target to allow students to reproduce the results from E. coli UvrD as a control in parallel with novel Taq UvrD results. This approach creates a controlled environment for students to practice the critical thinking, analysis, and experimental design skills necessary for a career in research. Students are given freedom in other aspects of experimental design, which incorporates in-class problem solving and guided literature review. For example, instead of telling students which restriction enzymes to use for cloning or analysis, students select from an array of available restriction enzymes based on their interpretation of plasmid maps. This procedure has the added benefit of expanding the research space because student groups will be generating different results based on their choices.

Each semester begins with the same sequence to expose each cohort to the core skills. Students clone a plasmid containing the UvrD into plasmid amplification and protein expression strains of E. coli. They perform restriction enzyme digests to confirm appropriate plasmid size. Students submit the plasmids for sequencing and reconstruct the plasmid from the results to confirm the gene sequence is present and correct. Following protein expression, the students purify UvrD using ammonium sulfate precipitation and affinity chromatography.

Once students have purified UvrD, the assays and experiments differ based on the outcomes of previous semesters. For example, at the outset of the course, students performed DNA unwinding assays using fluorescently labeled DNA and polyacrylamide gel electrophoresis. In more recent semesters, students have performed Förster resonance energy transfer to monitor DNA unwinding²⁵. These project-specific skills, while different between cohorts for some experiments, introduce an array of modern techniques that prepare students for a career in research.

Defining characteristics of this CURE

Longitudinal design

A graduate-level research project generally cannot be completed in a single semester. This CURE project is structured accordingly to span multiple semesters, where each semester has a new cohort of students whose work will build upon the efforts of the previous cohort. For example, because Taq is a thermophile, Taq UvrD is likely more active at high temperatures. DNA melting is also a function of temperature. Students performed assays to determine the maximum temperature of DNA substrate stability for unwinding assays, results of which were incorporated in lab procedures. The longitudinal course structure allows pursuit of more impactful research goals than ones designed to be performed start to finish in one or two semesters. Furthermore, a long-running project is more representative of research that occurs in academic and industrial laboratories.

One of the main challenges of this method is that in some semesters, students spend much of their time overcoming experimental failure and troubleshooting. Although such experiences can discourage students, they more aptly prepare students for the realities of scientific research. Furthermore, they provide the opportunity to teach students the roles of negative data and method development in the scientific process²⁰. This CURE structure results in an asymmetrical curriculum between semesters, where cohorts are trained in “project-specific skills” that differ between semesters. The core skills, however, are taught every semester such that all students are exposed to those techniques.

Larger student groups with diverse levels of experience

Pre-course surveys are utilized to determine students’ experience levels. Prior to the start of a semester, students respond to a questionnaire regarding confidence and prior experience with the core bench skills taught in this CURE and general research skills like literature review and data analysis. Through this assessment, experience is evenly distributed within student groups. Less experienced students are grouped with more experienced students to provide built-in peer mentoring. This group organization provides benefits for students who may struggle with the material^26,27 and helps more experienced students identify knowledge gaps through discussion²⁸. Harnessing peer mentoring also effectively increases the instructor-to-student ratio, allowing TAs to focus on students who need the most help.

Biochemical research often relies on collaborative group contributions for benchwork and research products (reports, presentation materials, etc.). The scope of experiments, specifically running parallel controls and the novelty of some techniques, could potentially require more time than allotted for some lab sessions, leading to incomplete experiments or slower progression. Accordingly, students work in groups of four rather than the traditional two partners. The group structure facilitates cooperative progression through lab activities, promotes collaborative writing and peer review, and efficient performance of experimentation. Students are expected to divide work duties amongst themselves, reducing the setup time for the more laborious experiments. Novel research necessitates troubleshooting and optimization, and groups can be divided to run multiple experimental conditions simultaneously. Additionally, each group can be assigned a different set of variables to test, expanding the breadth of research that can be covered in a class period.

To ensure equal contribution from group members, participation scores are awarded for each lab period and a significant component of collaborative writing grades are reserved for individual contributions. Additionally, quizzes, written exams, pre-lab activities, and laboratory notebooks are assessed on an individual basis. Finally, each group member is expected to present a portion of the final poster and field questions from instructors and colloquium attendees.

Lab meeting-style recitation

A vital part of the CURE is weekly recitation, where students analyze and interpret results from previous experiments, create hypotheses, and design experiments with instructor feedback. Emphasis is placed heavily on data analysis and interpretation because the progression of the project is dependent on the outcomes of prior experiments. Performing group analysis allows students to openly discuss their findings with one another. These group analyses, in line with the expectations of professional research, provide “soft skills” such as critical thinking and data interpretation that are often less developed in early-career scientists than physical bench skills.

Each group performs analysis of the previous week’s results and presents their conclusions and suggestions for the next experimental step in a lab meeting-style discussion. If the experiments were determined to have failed, the students hypothesize explanations for the outcomes and experimental remedies. This environment allows the instructors to introduce analysis techniques, software, and conceptual information. For example, groups are asked to annotate figures from recent experiments and describe their interpretation of the outcome. Students must use the data to defend those interpretations within their group, between groups who tested similar conditions, or to the entire class, while instructors provide feedback in real time. In addition, students participate in “elevator talks” throughout the semester, which requires the students to refine their understanding, such that they can concisely communicate their results.

Presentation of Experimental Results

In professional science, communication of results is necessary; however, few students reported having experience presenting the findings of their own research. To overcome this lack of training, students are required throughout the semester to present their findings in multiple formats.

Individually written lab reports throughout the semester have been replaced by a single, collaborative written report for each group. This approach more closely resembles real-world scientific writing, where multiple scientists contribute to a single body of work, a process in which students likely have not previously participated. The student groups are responsible for member accountability. Students submit sections of the written report following completion of semester milestones (e.g., completion of plasmid cloning or protein purification) so that students receive incremental feedback, allowing them to refine their knowledge and writing skills. Notably, this experience more closely resembles the iterative and collaborative process of writing scientific papers.

At the completion of the course, groups are required to present a poster at the UNC-wide undergraduate research symposium. Creating a poster based on their semester research reinforces students’ knowledge and introduces graphic design skills for effective data visualization. The abbreviated presentation format also provides experience in distilling essential information from their results. This activity is facilitated by elevator talks throughout the semester. The research symposium also resembles the format utilized at conferences that many students will attend in their future careers. For many students, this is their first experience in presenting an original research project to an external audience, and it provides the students with ownership of their project.

Outcomes

Survey data

Post-course surveys are employed to gauge the effectiveness of this CURE implementation, including measures adapted from the Laboratory Course Assessment Survey (LCAS)²⁹ and Project Ownership-Content scale³⁰ (referred to herein as Ownership), both of which have been used to differentiate CUREs from traditional courses^31–33, as well as a follow-up to the skill confidence items from the pre-course survey. The LCAS contains three subscales measuring students’ perceptions of CURE features: Collaboration, Discovery, and Iteration. Collaboration (6 items) addresses the frequency with which students are encouraged to work with instructors and classmates. Discovery (5 items) describes the degree to which students experienced opportunities to generate new knowledge in the discipline. Iteration (6 items) addresses the amount of revision and repetition students used to solve problems and validate their results. While the LCAS measures students’ perceptions of course attributes, Ownership focuses on students’ sense of engagement and agency toward the project. Full survey prompts and response scales are displayed in Table 1. Additionally, two open-ended reflection questions seek to identify what course components students found most challenging and rewarding.

Table 1.

Post-semester survey prompts by category with abbreviations and scale

Category	Label	Prompt	Scale
Collaboration	C1	I was encouraged to discuss elements of my investigation with classmates or instructors.	1 = Never 2 = One or Two Times 3 = Monthly 4 = Weekly
	C2	I was encouraged to reflect on what I was learning.
	C3	I was encouraged to contribute my ideas and suggestions during class discussions
	C4	I was encouraged to help other students collect or analyze data.
	C5	I was encouraged to provide constructive criticism to classmates and challenge each other’s interpretation.
	C6	I was encouraged to share the problems I encountered during my investigation and seek input on how to address them.
Discovery	D1	I was expected to generate novel results that were unknown to the instructor and that could be of interest to the broader scientific community or others outside of class.	1 = Strongly Disagree 2 = Disagree 3 = Somewhat Disagree 4 = Somewhat Agree 5 = Agree 6 = Strongly Agree
	D2	I was expected to conduct an investigation to find something previously unknown to myself, other students, and the instructor.
	D3	I was expected to formulate my own research questions or hypothesis to guide an investigation.
	D4	I was expected to develop new arguments based on data.
	D5	I was expected to explain how my work has resulted in new knowledge.
Iteration	I1	I was expected to revise or repeat work to account for errors or fix problems.	1 = Strongly Disagree 2 = Disagree 3 = Somewhat Disagree 4 = Somewhat Agree 5 = Agree 6 = Strongly Agree)
	I2	I had time to change the methods of the investigation if it was not unfolding as predicted.
	I3	I had time to share and compare data with other students.
	I4	I had time to collect and analyze additional data to address new questions or further test hypotheses that arose during the investigation.
	I5	I had time to revise or repeat analyses based on feedback.
	I6	I had time to revise drafts of papers or presentations about my investigation based on feedback
Ownership	O1	My research will help to solve a problem in the world.	1 = Strongly disagree 2 = Disagree 3 = Neither agree nor disagree 4 = Agree 5 = Strongly agree
	O2	My findings were important to the scientific community.
	O3	The research question I worked on was important to me.
	O4	My research project was interesting.
	O5	My research project was exciting.
	O6	I faced challenges that I managed to overcome in completing my research project.
	O7	I was responsible for the outcomes of my research.
	O8	The findings of my research project gave me a sense of personal achievement.
	O9	I had a personal reason for choosing the research project I worked on.
	O10	In conducting my research project, I actively sought advice and assistance.

To avoid confounding the results with the effects of the COVID-19 pandemic, this report focuses on survey responses (n=70) collected in Fall 2017 (n=9), Spring 2018 (n=25), Fall 2018 (n=10), Spring 2019 (n=14), and Fall 2019 (n=12). Although we do not have survey data prior to implementing the CURE and cannot perform a direct comparison to the previous curriculum, mean overall and category-specific survey results comparable to those established in Corwin et al.²⁹ were observed (Figure S1), suggesting the outcomes are comparable with those from CUREs at other institutions. Based on pre- and post-semester surveys, confidence in all core skills significantly increased across all semesters (Figure 1), with the lone exception of agarose gel electrophoresis in Spring 2019 (p=0.053). This result confirms that students are gaining the necessary core skills in this research-based class.

Figure 1. Student confidence in core skills.

Students ranked their confidence with six skills pre- and post-semester: agarose gel electrophoresis, sodium-dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), bacterial plasmid transformation, restriction enzyme digestion of DNA, protein purification, and enzyme activity assays. Under each skill, three plots. Left: mean scores by semester for pre- (open circle) and post-semester (closed circle) surveys. Semesters are abbreviated as F/S (Fall/Spring) and the last two digits of the year. Error bars denote standard deviation. Increases in confidence denoted with an asterisk (*) were determined to be statistically significant using a paired Student’s t-test (p < 0.05). Center: histogram of pre- to post-semester changes for all students. Bars are colored according to movement direction (red = loss, yellow = no change, blue = gain) and increasing hue intensity corresponds to larger changes. Right: pie chart showing the percentage of students who reported a loss (red), gain (blue), or no change (yellow) in confidence.

Across all LCAS/Ownership categories, responses were generally strong, with the vast majority above the 75^th percentile of their respective scales (Figures S2–5). The only notable exception to this pattern was having a personal reason for choosing the project (O9), which scored markedly lower than all other questions across all semesters. These lower responses are understandable given that the instructors chose the research topic. Within each category, scores are similar from semester to semester (Figure 2a), suggesting students receive a consistent educational experience despite instructor and TA changes.

Figure 2. Post-semester student survey results.

A) Mean survey scores by semester for each of the four categories. Scales were normalized to the maximum possible score for each category (Collaboration, 4; Discovery, 6; Iteration, 6; Ownership, 5). A linear regression was performed in Python and is displayed as a solid line surrounded by the shaded 95% confidence interval. B) Open-ended reflection responses for most rewarding (left) and challenging (right) aspects of the course. Responses were inductively analyzed to identify recurrent themes and binarized, where each theme received a score if it was mentioned (1) or not (0). Each student response could contain multiple themes.

Survey responses generally correlate within categories

To examine relationships between survey items across semesters, Exploratory Factor Analysis (EFA) was employed using Python. A Bartlett test determined that enough structure existed in the data to perform dimensionality reduction (c² = 1959.75, p < 0.001)³⁴ and a Kaiser-Meyer-Olkin Test confirmed adequate sampling for EFA (0.71)^35,36. Factors were identified using a maximum likelihood approach, and Oblimin rotation was used because correlations between LCAS categories were previously shown²⁹.

Of the six factors identified with eigenvalues above 1³⁷, the lowest two (1.10, 1.04) fell within the low acceleration region of the scree plot (Figure S2) and were excluded³⁸, leaving a 4-factor solution (Table 2). A caveat of this analysis is the low subject to item ratio (2.6:1), so a loading cutoff of 0.5 was applied to limit results to the strongest associations. Every survey item loaded into a single factor, apart from two that failed to load in any factor (O3, O9). All factors displayed reliable internal consistency with Cronbach’s alpha values > 0.8^39–41. The four factors identified were Revision/Achievement, Collaboration, Discovery, and Troubleshooting/Community. The three LCAS categories — originally identified through EFA²⁹ — were separated, with Collaboration and Discovery each comprising distinct factors. Interestingly, the Iteration questions were divided between those involving progressive revision of work (I3 – I6) and troubleshooting (I1 – I2). Ownership items were split between those encompassing personal achievement (O4 – O8, O10) and community importance (O1, O2).

Table 2.

Factor loadings for survey items^a

	Revision/Achievement	Collaboration	Discovery	Troubleshooting/Community
C1	—	0.539	—	—
C2	—	0.800	—	—
C3	—	0.645	—	—
C4	—	0.691	—	—
C5	—	0.799	—	—
C6	—	0.901	—	—
D1	—	—	0.766	—
D2	—	—	0.795	—
D3	—	—	0.706	—
D4	—	—	0.546	—
D5	—	—	0.629	—
I1	—	—	—	0.579
I2	—	—	—	0.737
I3	0.716	—	—	—
I4	0.802	—	—	—
I5	0.653	—	—	—
I6	0.677	—	—	—
O1	—	—	—	0.656
O2	—	—	—	0.577
O3	—	—	—	—
O4	0.662	—	—	—
O5	0.581	—	—	—
O6	0.614	—	—	—
O7	0.811	—	—	—
O8	0.727	—	—	—
O9	—	—	—	—
O10	0.840	—	—	—
Cronbach’s α	0.93	0.88	0.85	0.88
Factor correlations
R/A	—
C	0.21	—
D	−0.41	−0.26	—
T/C	0.93	0.52	−0.32	—

^aFactor loadings below 0.5 omitted

Semester identifiers were removed from student responses to eliminate potentially identifiable information resulting from small cohort sizes, so semester mean values were examined to evaluate trends over time. Because Revision/Achievement and Troubleshooting/Community are comprised of items with different scales, item means were converted to Z-scores and used to generate factor means for each semester. Relative changes over time were examined by Pearson correlation and classified by magnitude for simplicity: strong (≥0.75), moderate (0.50 – 0.74), weak (0.25 – 0.49), and none (≤0.24). Troubleshooting/Community displayed a strong positive correlation with Revision/Achievement and a moderate positive correlation with Collaboration. Discovery displayed weak negative correlations will all factors.

To determine underlying causes of factor correlations, correlations between raw semester means for each item were calculated and hierarchically clustered using a pairwise spatial distance algorithm (Figure 3). The dendrogram (top) identifies two main groups, shown by the uppermost fork. The grouping on the left contains all Iteration and all but one Ownership question, explaining the strong positive correlation between Revision/Achievement and Troubleshooting/Community, which are comprised of Iteration and Ownership. Discovery questions are entirely in the second cluster due to negative correlations with most of the Iteration items, Ownership, and Collaboration items, resulting in negative correlations with the other factors. Interestingly, Collaboration is split evenly between the two clusters despite occupying a single, distinct factor. While items C1 and C4 are negatively correlated with Troubleshooting/Community items, the remaining four Collaboration items have positive correlations which drive the moderate positive correlation between those factors.

Figure 3. Correlations between survey items.

A Pearson correlation of the change in normalized mean semester scores for each question was performed in Python. The dendrogram (top) shows the hierarchical sorting. Increasing hue intensity represents stronger positive (blue) and negative (red) correlations. Color-coded boxes and question abbreviations bordering the correlations correspond to Table 1. Questions are paraphrased for readability.

Correlation clustering provides insight into course components

Smaller clusters of strong correlations provide insight into the relationship between course elements and student perceptions. Some relationships are predictable, such as the strong positive correlations between the group of prompts involving overcoming challenges (O6) through evolving methods (I2), repeating work (I1), and contributing suggestions to class discussions (C3). Other trends are less predictable. Strong positive correlations between the bulk of the Iteration (I3–6) and two Ownership questions (O1, O3) suggests that repeating, revising, and sharing work with others is associated with an increased sense of importance. Another block of strong correlations suggests that students who take personal responsibility for their work (O7) seek more advice (O10), and these students also tend to have a greater sense of personal accomplishment (O8) and contribution to the field (O2).

Effect of lab meeting-style recitation on survey responses

We equate items relating to class discussions, sharing ideas, and problem solving as responses associated with the lab-meeting style recitations and in-lab discussions. A network of mostly moderate to strong positive correlations connects the Iteration questions (statements reflecting the perceived amount of time for repeating and revising work) to the Ownership statements (except O4). The Collaboration statements mixed into this network involve concepts such as sharing issues (C6), contributing to discussions (C3), and reflecting upon the work (C2). Additionally, the Collaboration topic of constructive criticism (C5), while showing mostly weak to moderate negative correlations with the Iteration and Ownership group, did show moderate to strong positive correlations with aspects of the recitation associated with adapting and overcoming challenges (O6, I2, C3).

Although there are many positive correlations between Iteration and Ownership, there is an interesting pattern of moderate and strong negative correlations between the Iteration (I1–6) and Ownership statements (O1, O3, O5, O6) associated with the recitation structure and one specific collaboration question about discussing investigational elements (C1). These relationships are counterintuitive because recitations are typically based around group analysis and discussion. One possible explanation is that discussion may lead to generation of ideas for new experiments without ample time to complete those new experiments, leaving the students feeling somewhat unfulfilled. Failing experiments were consistently cited by students across semesters as a major hurdle, a relationship evident in the survey results as well. The strongest negative correlation (-0.93) is between two Ownership statements: overcoming challenges (O6) and interest in the project (O4). If students felt they did not overcome the challenges, specifically failing experiments, they may become disinterested or cynical. Indeed, moderate and strong negative correlations exist between the Iteration questions and personal interest in the project (O4). Together, these results suggest that student interest wanes if there is insufficient time to adequately address problems during the semester.

Emphasis on research products bolstered project ownership

Adequate time for poster and presentation work (I6) showed strong positive correlations with personal importance (O3) and solving a world problem (O1) and moderate positive correlations with research excitement (O5) and community importance (O2). Reflection responses corroborate these observations and indicate students found the creation and presentation of research posters to be among the most rewarding components of the course (Figure 2b). Several responses emphasized that students had not previously been asked to present results of original research. Together, these results indicate that the emphasis on original research products is filling an educational gap, enhancing course outcomes by imbuing a sense of importance in the project.

Scope of work and troubleshooting affect project ownership

One of the most cited challenges in the student reflections was course design and instructor communication (Figure 2b), with many mentions occurring during the second semester of the CURE implementation (Spring 2018). There are noticeable score reductions across the Iteration questions, where the prompts address time constraints (Figure 2a, Figure S5). Most Ownership responses also decreased (O1–2, O4–8, O10, Figure S6), suggesting the Spring 2018 students had less of a sense of accomplishment after the course than students in other semesters. During this semester, mutant enzyme design was added in addition to troubleshooting experiments from the prior semester, and the scope of experiments may have been overambitious. This observation suggests that the scope of work each semester must be carefully chosen to not overwhelm students.

The Discovery prompts, generally statements of student perceptions of the novelty and importance of the research, were mostly characterized by negative or neutral correlations with the Iteration and Ownership statements. This relationship is understandable because instructors guided the students in troubleshooting experiments to obtain the hypothesized outcomes. Two particularly strong negative correlations between project excitement (O5) and explaining new arguments (D4) and knowledge (D5) suggest that the more students reflected on the novelty of their results, the less exciting the project was to them. Those feelings surface in the student reflections (Figure 2b), where some students specified a lack of sufficient troubleshooting time as a major challenge, while others describe the reward of experimental success in the context of troubleshooting.

Takeaways

The CURE teaches the same skills as early graduate work

The first year of graduate school is often spent gaining experience and confidence with bench skills and soft skills such as understanding literature, collaboration, analyzing data, generating hypotheses, and presenting original research. This course was designed around these characteristics of early graduate work, and these results suggest that the CURE is providing students with meaningful experience in those skills. Generally, student confidence in core laboratory skills increased across all semesters. Additionally, the standard deviations uniformly became smaller following completion of the course, suggesting that the knowledge gap between inexperienced students and those with prior extracurricular lab experience was reduced.

Collaboration and data analysis are emphasized in this curriculum. Student reflection responses included these major points of emphasis as both challenges and rewards, and these sentiments also appear in survey scores. The absence of these skills in the previous curriculum was a major motivator to update the course, and the results suggest that incorporation of lab meeting-style recitations, research posters, and presentations has exposed the students to these concepts, in some cases for the first time.

One area that could be emphasized more in the future is independent reading and interpreting literature. The course materials are provided as a mixture of literature sources and protocols, but to this point students have not been required to perform any independent query or analysis, but rather have been led in journal club-style group discussions. Additionally, measures, such as providing practical or newsworthy applications of similar studies, must be incorporated to ensure all students are engaged in the recitations, which could have a positive effect on project ownership.

Established controls are a necessity

One challenge in this approach is the unpredictable nature of research, and semesters may end without resolving experimental issues. The results suggest students are somewhat demoralized when semesters are consumed by troubleshooting. To avoid such situations, establishing reliable control experiments should be emphasized. Solid controls not only provide experimental validation but also act as fail-safes for unpredictable experiments. For example, we have modified the experiments to incorporate commercially available UvrD homologs which have undergone rigorous quality assurance measures to guarantee activity, providing a level of certainty while testing hypotheses with unknown outcomes.

Communication is key to a successful CURE implementation

An often-cited challenge in the course was communication, particularly in the early semesters. Students had trouble writing without rubrics and were confused about which experiments would be performed prior to each lab session. These communication issues arose from our inexperience with this curriculum and the novel nature of the project. As the course has evolved, the scope of work has been adjusted when too ambitious. New controls and alternative methods have also been adopted, as would happen in professional research, and instructor communication with students has become a focal point. Additionally, although a strict rubric for writing assignments is not provided because it does not accurately reflect scientific writing outside of the classroom, students are given more example material to guide them as they write their reports.

Communication goes beyond providing the students with more information about lab periods and research projects. Analysis of the survey data provides insight into ways messaging can be improved and change students’ perceptions of this course and scientific research. For example, Discovery questions – those questions dealing with generating novel results – generally correlated negatively with Iteration and Ownership questions. Our interpretation is that students saw the troubleshooting work as an impediment to the research goal of finding something new; however, determining the proper procedures and conditions for performing experiments is often the hardest part of research and the results of those troubleshooting experiments are valuable to the community.²⁰ This assessment was incorporated in real time, providing an information session during a recitation period where these points were discussed with the students to emphasize that the results of failed experiments are still valuable knowledge. Additionally, after the communicated project goals have been expanded beyond a novel biochemical result to include pioneering method development.

Overall, these results demonstrate that this CURE provides an undergraduate course that better prepares students for the rigors of a career in scientific research. The course is still evolving and working towards publishing the combined work of multiple student cohorts. We hope that this account of this CURE implementation, including our interpretations of what went well and what can be improved, can help educators at institutions of every level develop their own programs to provide research experience to those students who might otherwise not have any. It is our hope that providing realistic expectations for students entering the field will have a positive impact on students’ mental health and self-confidence, as well as providing better-prepared entry-level candidates for all sectors of the scientific community.