Exaptation: Academic mentees' career pathway to be independent and impactful

Abstract

In science, mentees often follow their mentors' career paths, but exceptional mentees frequently break from this routine, sometimes even outperforming their mentors. However, the pathways to independence for these excellent mentees and their interactions with mentors remain unclear. We analyzed the careers of over 500,000 mentees in Chemistry, Neuroscience, and Physics over the past 60 years to examine the strategies mentees employ in selecting research topics relative to their mentors, how these strategies evolve, and their resulting impact. Utilizing co-citation network analysis and a topic-specific impact allocation algorithm, we mapped the topic territory for each mentor-mentee pair and quantified their academic impact accrued within the topic. Our findings reveal mentees tend to engage with their mentors' less-dominated topics and explore new topics at the same time, and through this exaptive process, they begin to progressively establish their own research territories. This trend is particularly pronounced among those who outperform their mentors. Moreover, we identified an inverted U-shaped curve between the extent of topic divergence and the mentees' long-term impact, suggesting a moderate divergence from the mentors' research focus optimizes the mentees' academic impact. Finally, along the path to independence, increased coauthorship with mentors impedes the mentees' impact, whereas extending their collaboration networks with the mentors' former collaborators proves beneficial. These findings fill a crucial gap in understanding how mentees' research topic selection strategies affect academic success and offer valuable guidance for early-career researchers on pursuing independent research paths.

Figures (5)

• Figure 1: Illustration of a mentor-mentee co-citation network and topic-specific impact measurement.a. The co-citation network contains all the papers published by a mentor-mentee pair. Each node represents a paper: circles for mentee papers, triangles for mentor papers, and overlapping shapes for mentor-mentee collaborated papers. Links connect nodes if the corresponding papers are co-cited by at least one common paper. Node colors denote the different communities identified by the fast unfolding algorithmblondel2008fast, and node sizes correspond to their impact, as determined by the algorithm described in panel c. b. The time series tracks the network's evolution, depicting the growth of papers by the focal mentee (upper half) and the pair-wise mentor (lower half) throughout their careers. Each shaded area's color matches a topic community in the co-citation network, with the height at each year point indicating the number of papers published on that topic. The vertical black line represents the year the mentor started supervising the mentee. c. The heuristic algorithm quantifies the impact of a focal mentee by analyzing co-citations within each topic community. The mentors' papers ($r_{j,i}$) and mentees' papers ($e_{j,i}$) are organized by their topic communities in the left column, color-coded to align with the communities $j$ identified in panel a. Squares in the right column signify papers that co-cite any of the mentor-mentee papers, with colors matching the communities of mentor-mentee papers. Solid lines connect papers with the same colors, showing co-citations within communities, while dashed lines indicate cross-community co-citations.
• Figure 2: Interaction patterns of research topics between mentor and mentee.a. Illustration of three topic selection strategies for mentees. In each toy co-citation network, circles represent mentee papers, and triangles represent mentor papers. Nodes within the same topic community are enclosed within a dashed circle. We identify three distinct topic selection strategies: "Pure follow," where mentees focus exclusively on their mentor’s topics; "Follow & innovate," where mentees work on both their mentor’s topics and explore new areas; and "Pure innovate," where mentees fully diverge to pursue topics independent of their mentor’s work. The values of $R$, defined as the ratio of the number of the mentee's new topics among all their topics, have been marked for each toy network. b-d. The complementary cumulative distribution function (CCDF) of the ratio of new topics ($R$) for mentees. The purple stepped line represents the CCDF ($R$) for all mentees in our dataset, while the yellow stepped line indicates the CCDF ($R$) for elite mentees, defined as those in the top 20% by cumulative citations. Inset: the fraction of mentor-mentee topic interaction patterns in panel a for all (purple bars) and elite mentees (yellow bars). e-g. The average number of topics of different types (primary, secondary and new) pursued by all mentees and elite mentees over their careers (see the distribution of the number of mentees' topics across different types in Supplementary Fig. S5). h-j. The average ratio of impact across different topic types for all mentees and elite mentees over their careers. This measure is computed and compared separately for all mentees and elite mentees with two topic selection strategies, namely, "follow & innovate" and "pure follow".
• Figure 3: Evolution of cumulative impact across different topic types for all mentees (a-c) and outperforming mentees (d-f). The solid black line represents the average cumulative impact earned from mentees' papers, while the dashed black line shows the average cumulative impact of their corresponding mentors' papers over the career years since their first publication. The colored lines—orange for primary topics, green for secondary topics, and grey for new topics—track the average cumulative impact evolution for mentees' papers that are dedicated exclusively to these specific topic areas. The insets display boxplots for the ratio of the number of mentees' papers by topic type for each decade, with blue triangles denoting the mean and purple lines indicating the median number of papers.
• Figure 4: Relationship between the likelihood of mentees outperforming their mentors and their research topic similarity.a-c. Impact difference between mentees and their paired mentors in primary and secondary topics. The horizontal axis shows the impact difference for primary topics and the vertical axis for secondary topics, calculated by subtracting the mentee's impact from the mentor's. Yellow points indicate outperforming mentees, and blue points depict all mentees. The insets, one per quadrant, report the percentage of points that fall into each quadrant for the respective group of mentees distinguished by the color of the nodes. d-f. The positions of mentees in the primary–secondary–new topic triangle are determined by how many relative citations each mentee accrued from the respective topic type. The circle size denotes the total number of citations received by mentees, and the color denotes the group of mentees: yellow for outperforming mentees; and purple for all mentees. g-i. The inverted-U relationship between the topic similarity of mentees and mentors and mentees' cumulative impact. Here, the topic similarity is measured by the average length of the shortest path between the mentor and mentee nodes in the co-citation network, as calculated by the formula \ref{['eq:ave_shortest_path']}. Each blue point represents the average cumulative impact relative to the average path length (x-axis), with error bars showing the standard error. The blue line with shading represents the quadratic fitting curve of all the scatter points in the inset, which depicts the relationship between the average distance to their mentor's papers in the co-citation network and the mentee's cumulative impact.
• Figure 5: The regression analysis of the association between the average network distance and the mentees' cumulative impact in Chemistry (a), Neuroscience (b) and Physics (c). Results from field-specific non-linear regressions ($R^2$ = 0.202, 0.203, 0.162 for Chemistry, Neuroscience, and Physics, respectively, Model 10 in Supplementary Table S6-S8), whose dependent variable is the mentees' cumulative impact. Note that the coefficients (grey nodes) for colla_work_count are derived from Models 7 presented in Supplementary Table S6-S8. The statistical significance of the variables is presented at the left of each value (* p < 0.05; **p < 0.01; ***p < 0.001).