Tissue preservation can affect geometric morphometric analyses: a case study using fish body shape

Abstract

In geometric morphometrics, the extent of variation attributable to non-biological causes (i.e. measurement error) is sometimes overlooked. The effects of this variation on downstream statistical analyses are also largely unknown. In particular, it is unclear whether specimen preservation induces substantial variation in shape and whether such variation affects downstream statistical inference. Using a combination of empirical fish body shape data and realistic simulations, we show that preservation introduces substantial artefactual variation and significant non-random error (i.e. bias). Most changes in shape occur when fresh fish are frozen and thawed, whereas a smaller change in shape is observed when frozen and thawed fish are fixed in formalin and transferred to ethanol. Surprisingly, we also show that, in our case, preservation produces only minor effects on three downstream analyses of shape variation: classification using canonical variate analysis, permutation tests of differences in means and computations of differences in mean shape between groups. Even mixing of differently preserved specimens has a relatively small effect on downstream analyses. However, we suggest that mixing fish with different preservation should still be avoided and discuss the conditions in which this practice might be justified.

Optimizing digitalization effort in morphometrics

“…Finally, when there is a doubt on repeatability and when datasets do not overlap between users, it is also possible to remove the effect of the origin of the data by introducing this as a factor in the analysis (e.g. [ 16 ]) but this requires the critical assumption that the variation of the hypothetical bias is not in the same direction as the biological variation explored. This can however be done in trying to keep balanced data among data providers for avoiding violating “the marginality principle” in linear modeling.…”

“…While some studies have evidenced small variation between multiple operators when compared with the targeted biological signal (e.g. [ 16 ]) others have demonstrated that IO bias can lead to substantial variation on geometric morphometric analyses (e.g. [ 17 ]).…”

“…[ 17 ]). IO measurement errors (MEs) can have dramatic impacts on morphometric studies, especially when the phenotypic variation under investigation is subtle or varies in the same direction as the explored biological variation [ 16 ]. This is true particularly as the methods such as geometric morphometrics (using coordinates of landmarks, curves, or surfaces) are especially used for detecting small-scale shape and size variation [ 18 , 19 ].…”

“…Several approaches have been proposed to assess error in morphometrics based on replicated measurements [ 12 , 20 , 21 ]. Less studies have addressed the problem of pooling data from various sources [ 13 , 14 , 16 ].…”

Are geometric morphometric analyses replicable? Evaluating landmark measurement error and its impact on extant and fossil Microtus classification

“…Quantification of presentation and projection error requires replication which generally must be conducted at specimen housing facilities, and is seldom assessed in the literature (but see Fruciano, ; Fruciano et al, ; Robinson & Terhune, ) though their potential for obscuring biologically meaningful shape variation is considerable (Fruciano, ; Zelditch et al, ). Few studies demonstrate how several of these error types can combine to impact statistical result‐based inferences (but see Fruciano et al, ,; Robinson & Terhune, ; Vergara‐Solana, García‐Rodríguez, & Cruz‐Agüero, ). This context is important because ecological, archeological, and paleontological studies often use statistical grouping analyses (e.g., linear discriminant analysis (LDA)/canonical variate analysis) to determine the taxonomic or ecological affinity of unknown specimens (Kovarovic, Aiello, Cardini, & Lockwood, ; Webster & Sheets, ).…”