Search bioRxiv⌕ Search
Home/Archive/bioRxiv

bioRxiv · 10.1101/2025.02.08.637263

From citizen to scientist: evaluating ant observation accuracy on a Mediterranean archipelago

Arcos, J.
View original source ↗Export metadata ↓

Abstract

Citizen science platforms, such as iNaturalist, have become invaluable tools for biodiversity monitoring, allowing non-experts to contribute photo-based observations that are then identified by the user community. However, one of the major challenges of these platforms concerns the accuracy of these identifications, especially for taxa requiring specialized knowledge or more sophisticated means of identification. In this study, I assess the reliability of ant identifications in the Balearic Islands by comparing community-generated identifications with expert validations across multiple taxonomic levels. Based on 300 iNaturalist observations, I analyze the influence of user experience, AI-generated suggestions, image quality, and community participation on identification accuracy. The results indicate that species-level accuracy was 69.91%, increasing to 90.91% at the genus level and 99% at the family level. Experienced users significantly improved accuracy, while early-stage identifications at finer taxonomic resolutions increased the likelihood of correct consensus. AI-assisted identifications performed well for frequently recorded species but struggled with underrepresented taxa. Surprisingly, photo quality had minimal impact, as common species were often identifiable even from low-resolution images. Additionally, the dataset documented exotic species, including the first record of the Formica rufibarbis complex in Mallorca. These findings highlight both the strengths and limitations of citizen science in taxonomic research and emphasize the need for strategies to enhance data reliability for conservation and biodiversity monitoring.

Explore related subjects

zoologyPreprint

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Arcos, J.. 2025-02-09. From citizen to scientist: evaluating ant observation accuracy on a Mediterranean archipelago. https://doi.org/10.1101/2025.02.08.637263

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Late lactation represents the main window for sow-to-piglet transmission of persistent gut strains

The gut microbiota plays a key role in piglet health, and maternal microbial transmission may represent a promising lever to shape early-life microbiota and prevent post-weaning digestive disorders. This study aimed to better characterize sow-to-piglet microbiota transmission and persistence using a long-read metabarcoding approach targeting the 16S-ITS-23S region. Fecal samples (n = 204) were collected from 17 families, a family being as sow and three of her piglets, at multiple stages: late gestation (G110), early (L6) and late lactation (L28) for sows; early lactation (L6), late lactation (L28), and 5 days post-weaning for piglets. To approximate strain-level resolution, a putative strain (PS) approach was developed by clustering ASVs (n = 6064) affiliated with the same species based on abundance covariance (r > 0.9), resulting in 4857 PS. Piglet microbiota progressively diversified during lactation and converged toward that of sow. In sows, 27 {+/-} 6% of PS were persistent from late gestation to late lactation. In piglets, only 4.2 {+/-} 2.5% of PS persisted from d6 to 5 days post-weaning. Persistent PS in piglets were mainly affiliated with Limosilactobacillus reuteri and Lactobacillus amylovorus followed with Holdemanella porci and H. biformis, Lentihominibacter hominis and Dorea formicigenerans. Shared PS were significantly higher within families than between unrelated pairs (p < 0.05). Maternal transmission peaked at the end of lactation (35 {+/-} 7% at L28). Persistent transmitted PS represented 2.7 {+/-} 1.6% (d6-post-weaning) and 15.4 {+/-} 5.6% (d28-post-weaning). Early-transmitted persistent PS were mainly affiliated with Limosilactobacillus reuteri, Lactobacillus amylovorus, and Paraeggerthella hominis, whereas late-transmitted persistent PS were associated with Prevotella spp., Sphaerochaeta globosa, and Bariatricus comes. These findings highlight the significance of maternal transmission in shaping the post-weaning microbiota and identify late lactation as a critical window for microbiota transfer.

zoology↗
bioRxiv2026

RISC-Bound Small RNA Sequencing Provides Insights into Guide Strand Selection and siRNA Trimming and Tailing Following Insecticidal dsRNA Delivery

RNA interference (RNAi) offers a sequence-specific approach to pest control. In insects, Dicer-2 processes double-stranded RNA (dsRNA) into small interfering RNA (siRNA) duplexes, from which the RNA-induced silencing complex (RISC) retains a guide strand. Only antisense-loaded RISC can mediate cleavage of the target transcript. However, how sequence features shape the RISC-bound siRNA pool in pests remains poorly understood, limiting opportunities for sequence optimization. Here, we profiled RISC-bound siRNAs following injection of 34 insecticidal dsRNAs targeting 11 essential genes in Tribolium castaneum larvae. We computationally reconstructed 7,879 siRNA pairs and examined associations between sequence features and strand bias. Differences in GC identity at terminal paired positions 1 to 5, used as a proxy for local thermodynamic asymmetry, correlated with strand bias, with the strongest correlations at the first two paired positions. ORF targeting and reduced predicted antisense self-folding were also associated with higher antisense fractions. Analysis of non-templated terminal additions revealed predominantly 3-prime uridylation, a known signature of small RNA turnover, along with putative 3-prime trimming. Among ORF-associated siRNA pairs, sense strands showed higher relative U-tailing abundance, based on 3-prime uridylated and putatively trimmed-and-3-prime-uridylated reads relative to perfect 21-nt reads, than antisense strands. Antisense strands with the least predicted self-folding also showed low relative U-tailing abundance. These observations are consistent with sequence-dependent contributions from both guide-strand selection and differential post-RISC-loading siRNA retention, although a causal link remains to be established. The identified associations provide a basis for testing whether dsRNA sequence optimization can improve pest control efficacy and reduce off-target activity.

zoology↗
bioRxiv2026

The second life of collection - how the scientific value of natural history holdings can be deliberately increased

Natural history collections cannot be assessed by the size of their holdings, because their scientific value is a multidimensional and dynamic property that can be deliberately shaped. The authors argue that this value can be systematically increased through purposeful curatorial and research actions, both without acquiring new material and through its continuous replenishment. We propose an analytical framework for describing the scientific value of a collection, based on five dimensions: epistemic distinctiveness, data integrity and accessibility, documented and potential research applications, unique taxonomic status, and compliance with ethical and legal standards. The framework is complemented by a dynamic perspective describing the conditions under which the value of a collection rises and falls over time. We then present a typology of actions that increase collection value incrementally, distinguishing four categories according to the object of the action: (i) enriching the data record, (ii) enriching the physical specimen, (iii) establishing formal status, and (iv) increasing visibility and interoperability. Each category of actions is linked to the corresponding dimensions of scientific value, so that the typology and the framework together form a single analytical tool. Both perspectives are illustrated with two collections deposited in the Natural History Collections of Adam Mickiewicz University in Pozna[n]. The first is an acarological collection of soil samples together with its derivatives, comprising collections of nomenclatural types, microscope slides, SEM images and molecular sequences deposited in global gene banks. The second is a conchological and virtual collection of the Roman snail (Helix pomatia), which illustrates the transition from a set of physical specimens to a resource of spatial digital data. The proposed approach provides practical tools for planning collection development, for communicating the importance of collections to funding bodies, and for identifying the untapped research potential of existing holdings.

zoology↗