The estimation of the covariance matrix is a key concern in the analysis of longitudinal data. When data consists of multiple groups, it is often assumed the covariance matrices are either equal across groups or are completely distinct. We seek methodology to allow borrowing of strength across potentially similar groups to improve estimation. To that end, we introduce a covariance partition prior which proposes a partition of the groups at each measurement time. Groups in the same set of the partition share dependence parameters for the distribution of the current measurement given the preceding ones, and the sequence of partitions is modeled as a Markov chain to encourage similar structure at nearby measurement times. This approach additionally encourages a lower-dimensional structure of the covariance matrices by shrinking the parameters of the Cholesky decomposition toward zero. We demonstrate the performance of our model through two simulation studies and the analysis of data from a depression study. This article includes Supplementary Material available online.

The global partition management software market is experiencing robust growth, driven by the increasing adoption of cloud computing, virtualization, and the expanding need for efficient data management across diverse organizational structures. The market, estimated at $2.5 billion in 2025, is projected to exhibit a Compound Annual Growth Rate (CAGR) of 12% from 2025 to 2033. This growth is fueled by several key factors. The rise of large-scale data centers necessitates sophisticated partition management tools for optimal resource allocation and performance. Simultaneously, the growing prevalence of hybrid cloud environments and the need for seamless data migration across platforms are creating significant demand for versatile and reliable software solutions. Furthermore, small and medium-sized enterprises (SMEs) are increasingly adopting these tools to improve data organization and simplify IT management tasks, contributing to market expansion. The web-based segment is currently the leading contributor to market revenue, owing to its accessibility and cost-effectiveness, while the cloud-based segment is anticipated to demonstrate the highest growth rate during the forecast period due to its scalability and enhanced security features. Geographic expansion into rapidly developing economies in Asia Pacific and the Middle East & Africa also contributes to the overall market expansion. However, certain restraints affect market growth. These include the high initial investment costs for advanced software solutions, the availability of free open-source alternatives, and the complexity of managing partitions in heterogeneous environments. Despite these challenges, the ongoing digital transformation across industries and the increasing reliance on data-driven decision-making ensure that the partition management software market will maintain a positive trajectory in the coming years. The market is witnessing a trend towards AI-powered automation in partition management tasks, improving efficiency and reducing the need for specialized IT personnel. Furthermore, vendors are focusing on enhanced user interfaces and improved integration capabilities with other enterprise software, furthering the market's evolution. The continued development of innovative features and solutions will drive substantial growth in the global partition management software market, creating considerable opportunities for industry players.

The global partition management software market is experiencing robust growth, driven by the increasing adoption of cloud computing, the proliferation of data storage devices, and the rising need for efficient data management across diverse platforms. The market, estimated at $2.5 billion in 2025, is projected to exhibit a Compound Annual Growth Rate (CAGR) of 12% from 2025 to 2033. This growth is fueled by several key trends, including the expanding enterprise data centers requiring sophisticated partition management solutions, the increasing prevalence of hybrid cloud environments demanding seamless data migration and partitioning capabilities, and the growing demand for enhanced data security and recovery features integrated within partition management tools. Furthermore, the market is segmented by application (large enterprises and SMEs) and type (cloud-based and web-based), with cloud-based solutions gaining significant traction due to their scalability and accessibility. The market's growth is, however, tempered by certain restraints. The high initial investment required for advanced partition management software can be a barrier to entry for smaller businesses. Furthermore, the complexity of some software solutions can pose a challenge for less technically proficient users. Despite these restraints, the long-term outlook for the partition management software market remains positive, driven by continuous innovation, the integration of advanced features like AI-powered data management and automation capabilities, and the rising demand for efficient data organization and control in an increasingly data-centric world. The competitive landscape is marked by a mix of established players and emerging companies, each vying to offer superior functionality, ease of use, and competitive pricing. This competitive pressure is expected to further accelerate innovation and benefit end-users.

The HapMap (haplotype map) projects have produced valuable genetic resources in life science research communities, allowing researchers to investigate sequence variations and conduct genome-wide association study (GWAS) analyses. A typical HapMap project may require sequencing hundreds, even thousands, of individual lines or accessions within a species. Due to limitations in current sequencing technology, the genotype values for some accessions cannot be clearly called. Additionally, allelic heterozygosity can be very high in some lines, causing genetic and sometimes phenotypic segregation in their descendants. Genetic and phenotypic segregation degrades the original accession’s specificity and makes it difficult to distinguish one accession from another. Therefore, it is vitally important to determine and validate HapMap accessions before one conducts a GWAS analysis. However, to the best of our knowledge, there are no prior methodologies or tools that can readily distinguish or validate multiple accessions in a HapMap population. We devised a bioinformatics approach to distinguish multiple HapMap accessions using only a minimum number of genetic markers. First, we assign each candidate marker with a distinguishing score (DS), which measures its capability in distinguishing accessions. The DS score prioritizes those markers with higher percentages of homozygous genotypes (allele combinations), as they can be stably passed on to offspring. Next, we apply the “set-partitioning” concept to select optimal markers by recursively partitioning accession sets. Subsequently, we build a hierarchical decision tree in which a specific path represents the selected markers and the homogenous genotypes that can be used to distinguish one accession from others in the HapMap population. Based on these algorithms, we developed a web tool named MAD-HiDTree (Multiple Accession Distinguishment-Hierarchical Decision Tree), designed to analyze a user-input genotype matrix and construct a hierarchical decision tree. Using genetic marker data extracted from the Medicago truncatula HapMap population, we successfully constructed hierarchical decision trees by which the original 262 M. truncatula accessions could be efficiently distinguished. PCR experiments verified our proposed method, confirming that MAD-HiDTree can be used for the identification of a specific accession. MAD-HiDTree was developed in C/C++ in Linux. Both the source code and test data are publicly available at https://bioinfo.noble.org/MAD-HiDTree/.

Marine communities undergo rapid changes because of human-induced ecosystem pressures. The Baltic Sea pelagic food web has experienced several regime shifts during the past century, resulting in a system where competition between planktivorous mesopredators is assumed to be high. While the two clupeids sprat and herring reveal signs of competition, the stickleback population has increased drastically during the past decades. Here, we investigate diet overlap between the three dominating planktivorous fish in the Baltic Sea, utilizing DNA metabarcoding on the 18S rRNA gene and the COI gene, targeted qPCR, and microscopy. Our results show niche differentiation between clupeids and stickleback and that rotifers play an important function in niche partitioning of stickleback, as a resource that is not being used, neither by the clupeids nor by other zooplankton. We further show that all the diet assessment methods used in this study are consistent but DNA metabarcoding describes the plankton-fish link at the highest taxonomic resolution. This study suggests that rotifers and other understudied soft-bodied prey may have an important function in the pelagic food web and that the growing population of pelagic stickleback is supported by the unutilized feeding niche offered by the rotifers.

1. Deterministic theories predict that local communities assemble from a regional species pool based on niche differences, thus by plant functional adaptations. We tested whether functional traits can also explain patterns in species dominance among the suite of co-occurring species. 2. We predicted that along a gradient of secondary succession the main driver of species dominance changes from environmental filtering in the relatively harsh (dry and hot) early successional conditions, towards increased competitive interactions and limiting similarity in later successional conditions (when light is limited). 3.We used the Kurtosis (K) (a measure of peakedness) of the functional trait distribution of secondary forest communities in high-diversity tropical rainforest in Chiapas, Mexico. The forests ranged 1 to 25 years in age, and we used 8 functional leaf traits related to a plants' carbon, water and heat balance. We calculated the functional trait distribution based on species dominance, where trait values were weighted by species' relative basal area, as well as based on species presence, all species counting once. "K-ratio" was subsequently computed by dividing kurtosis based on species dominance by kurtosis based on species presence. If the K-ratio is high, the dominant species are functionally similar and we interpreted this as environmentally driven functional convergence allowing species to become dominant. If the K-ratio is small, dominant species are a functionally dissimilar subset of the species present and we interpreted this as competitively driven functional divergence allowing species to become dominant. 4. We found that in early succession dominant species represent a functionally narrow subset of species with similar traits and in late succession dominant species increasingly represent a wide subset of the species present. This trend was found for traits that reflect photosynthetic performance and light capture, and indicates increased competition for light with succession. No trend was found for traits that indicate defense against herbivory, suggesting no successional changes in herbivore pressure. 5. Synthesis. This is one of the first studies showing that drivers of species dominance change along a gradient of secondary succession. During the early successional time window we evaluated, the importance of environmental filtering as a driving force fades away rapidly, and the importance of niche partitioning for species dominance starts to emerge.

This category of planning priorities in the CEC 2023 Land-Use Screens provides an estimate of terrestrial landscape condition based on the extent to which human impacts such as agriculture, urban development, natural resource extraction, and invasive species have disrupted the landscape across the State of California. It is based on the open-source logic modeling framework Environmental Evaluation Modeling System (EEMS) developed by Conservation Biology Institute (CBI). This multicriteria evaluation model result, last updated in 2016 and resolved at 1-kilometer square, spans values ranging from -1 to 1. The higher end of the spectrum indicates areas that are relatively intact based on the more than 30 input variables, and values in the lower end of the spectrum indicate where these human impacts to disturb the landscape and ecological function are relatively high.¹

In the adapted version of the CBI Terrestrial Landscape Intactness given here, the dataset is partitioned into high and low categories based on the mean. Values of the dataset that lie above 0.3 are considered highly intact and are used as an exclusion. Values of the dataset that are less than or equal to 0.3 are allowed to remain in consideration for resource potential. Applying the partition at the mean allows for lands that are relatively more intact than disturbed to be considered for resource potential. The high category of landscape intactness given by this dataset is used as an exclusion in both the Core and SB 100 Terrestrial Climate Resilience Study screens.

This layer is featured in the CEC 2023 Land-Use Screens for Electric System Planning data viewer.

More information about this layer and its use in electric system planning is available in the Land Use Screens Staff Report in the CEC Energy Planning Library.

[1] Degagne, R., J. Brice, M. Gough, T. Sheehan, and J. Strittholt. 2016. “Terrestrial Landscape Intactness 1 kilometer, California.” Conservation Biology Institute.https://databasin.org/datasets/e3ee00e8d94a4de58082fdbc91248a65/

Sources: Statistical Yearbooks, Newspaper-Articles, Scientific Publications.

This data describes the cellular metal concentrations of Phaeocystis antarctica and Cryothecomonas armigera following exposure to metals singly and in mixtures in laboratory studies. Microalgae were cultured in 80 mL of filtered (less than 0.45 um) seawater and low concentrations of nutrients supplemented with metal stocks to give a range of single and mixture exposures to the metals cadmium, copper, nickel, lead, and zinc. The cellular accumulation and partitioning are used to explain the metal's toxicity (cellular metal fractions are compared to the toxicity data provided in 10.4225/15/5ae93ff723ff8) and assess the risk bioaccumulation of metals to Antarctic marine microalgae may pose in the Southern Ocean food web.

The objective of this research was to estimate the elasticity of the demand for cocaine and heroin with respect to the price. Price elasticity is the percentage of change in the dependent quantity corresponding to a one-percent change in price. The project involved the development of an econometric model to determine price elasticity, given that national- and city-level data on the consumption of cocaine and heroin are insufficient or nonexistent. The researchers circumvented this lack of data by partitioning the desired elasticity into the product of two elasticities, involving a measurable intermediate quantity whose relationship to the quantity of consumption could be modeled and estimated by measurable techniques. The intermediate quantity used for this project was the fraction of arrestees testing positive for cocaine or heroin as measured by the Drug Use Forecasting (DUF) System. From the Drug Enforcement Administration's (DEA's) System to Retrieve Information from Drug Evidence (STRIDE) data, expected purity was computed by regressing on log quantity and dummy variables for location and quarter. Price series were produced by finding the median standardized price per expected pure gram for each location and quarter. Variables for Part 1, National Data, include year, quarter, standardized prices for a gram of cocaine and a gram of heroin, and expected purity of cocaine and heroin. The Cities Data, Part 2, cover city, year, quarter, number of observations used to compute the median price of cocaine and heroin, standardized prices, and expected purity.

Phylogenetic analysis of the spiders of the genus Cybaeulus, with outgroups in the marronoid clade. Data from six DNA markers, analyzed with maximum likelihood and parsimony.

PHYLOGENETIC ANALYSIS

We obtained sequences from 26 samples of the three known species of Cybaeolus, and of five additional species of Hahniidae. To these, we added legacy sequences of Cybaeolus and of other genera of Hahniidae, as well as representatives of the remaining families in the marronoid clade. For the new sequences, the extraction and amplification of DNA was made in the Laboratory of Molecular Tools at Museo Argentino de Ciencias Naturales (MACN), from tissues preserved in absolute alcohol at -18ºC. We targeted the markers histone H3 (H3), cytochrome oxidase subunit I (CO1), 28S ribosomal RNA (28S) and 16S ribosomal RNA (16S), previously used to estimate relationships of marronoid spiders (Wheeler et al., 2017). Details of extraction, primers and PCR protocols are the same as in Magalhaes & Ramírez (2022). Sequencing was outsourced to Macrogen Inc., South Korea. The resulting chromatograms were analyzed individually to detect contaminated sequences or ambiguous portions. In addition to these sequences obtained in the laboratory, we combined our data with additional sequences from previous work (Wheeler et al., 2017; Rivera-Quiroz et al., 2020), using the markers mentioned above plus 12S ribosomal RNA (12S) and 18S ribosomal RNA (18S). For the CO1 marker, additional sequences obtained by the Arachnology Division at MACN and deposited in the BOLDSYSTEMS platform (https://www.boldsystems.org/) were also used. Sequences were aligned with MAFFT Online v.7.463 (Katoh & Standley, 2013), using the L-INS-I algorithm. See Table 1 for list of vouchers and sequence identifiers.

Maximum likelihood
For the maximum likelihood analyses we used the program IQ-TREE 2.2.0 (Minh et al., 2020), partitioning the data by marker, and selecting the best combination of partitions and evolution models by Bayesian information criterion (best fitting models were TPM2+I+G4 for H3, GTR+F+I+G4 for 18S, GTR+F+I+G4 for 16S and 12S together, GTR+F+I+G4 for CO1, and GTR+F+I+G4 for 28S). Since the relationships of outgroup taxa in the resulting trees were slightly different to that found in recent phylogenomic studies, we used the study of Gorneau et al. (2023) based on ultraconserved elements as a backbone topology to constrain our tree search, considering only the taxa in common with our analysis (see supplementary Fig. S1); this means that all the rest of the taxa are free to move anywhere during tree search. Support for groups (branches) was estimated by 1000 cycles of ultrafast bootstrapping. Ten independent runs were performed; of those, six converged into nearly identical log likelihood values (-57417.7725 to -57417.9604) and identical topologies; the tree with top-ranking log likelihood is presented in Results, after collapsing branches with bootstrap below 0.5. To estimate the support of an alternative topology with Cybaeolus as sister to the rest of the hahniids, we used TNT 1.6 (Goloboff & Morales, 2023) to modify the optimal tree placing Cybaeolus in such position, and asked for the frequency of the branch of interest (all hahniids except Cybaeolus) in the 1000 bootstrapped trees previously saved by IQTREE.
Ancestral character states for the arrangement of spinnerets (grouped; separated in a transversal line) were estimated by maximum likelihood on the optimal tree, using the R packages phytools and ape, under the models ER and ARD, and the best fitting model selected by the Akaike information criterion.

Parsimony
For the parsimony analyses we used TNT 1.6. For the equal weights analysis, a heuristic search was made using a driven search with the default parameters of the “new technologies”, aiming for 10 independent hits to minimum length. The resulting trees were then submitted to an additional round of tree-bisection reconnection (TBR) branch swapping. These results were compared to a simpler search strategy of 300 random addition sequences, each followed by TBR, which produced 20 hits to minimal length. As both strategies reached the same trees with multiple independent hits, it is likely that the optimal trees were found. Finally, the strict consensus of all the optimal trees was obtained, and on this consensus the support values were calculated by means of 1000 bootstrap pseudoreplicates.

The Emotional Voice Messages (EMOVOME) database is a speech dataset collected for emotion recognition in real-world conditions. It contains 999 spontaneous voice messages from 100 Spanish speakers, collected from real conversations on a messaging app. EMOVOME includes both expert and non-expert emotional annotations, covering valence and arousal dimensions, along with emotion categories for the expert annotations. Detailed participant information is provided, including sociodemographic data and personality trait assessments using the NEO-FFI questionnaire. Moreover, EMOVOME provides audio recordings of participants reading a given text, as well as transcriptions of all 999 voice messages. Additionally, baseline models for valence and arousal recognition are provided, utilizing both speech and audio transcriptions.

Description

For details on the EMOVOME database, please refer to the article:

"EMOVOME Database: Advancing Emotion Recognition in Speech Beyond Staged Scenarios". Lucía Gómez-Zaragozá, Rocío del Amor, María José Castro-Bleda, Valery Naranjo, Mariano Alcañiz Raya, Javier Marín-Morales. (pre-print available in https://doi.org/10.48550/arXiv.2403.02167)

Content

The Zenodo repository contains four files:

EMOVOME_agreement.pdf: agreement file required to access the original audio files, detailed in section Usage Notes.

labels.csv: ratings of the three non-experts and the expert annotator, independently and combined.

participants_ids.csv: table mapping each numerical file ID to its corresponding alphanumeric participant ID.

transcriptions.csv: transcriptions of each audio.

The repository also includes three folders:

Audios: it contains the file features_eGeMAPSv02.csv corresponding to the standard acoustic feature set used in the baseline model, and two folders:

Lecture: contains the audio files corresponding to the text readings, with each file named according to the participant's ID.

Emotions: contains the voice recordings from the messaging app provided by the user, named with a file ID.

Questionnaires: it contains two files: 1) sociodemographic_spanish.csv and sociodemographic_english.csv are the sociodemographic data of participants in Spanish and English, respectively, including the demographic information; and 2) NEO-FFI_spanish.csv includes the participants’ answers to the Spanish version of the NEO-FFI questionnaire. The three files include a column indicating the participant's ID to link the information.

Baseline_emotion_recognition: it includes three files and two folders. The file partitions.csv specifies the proposed data partition. Particularly, the dataset is divided into 80% for development and 20% for testing using a speaker-independent approach, i.e., samples from the same speaker are not included in both development and test. The development set includes 80 participants (40 female, 40 male) containing the following distribution of labels: 241 negative, 305 neutral and 261 positive valence; and 148 low, 328 neutral and 331 high arousal. The test set includes 20 participants (10 female, 10 male) with the distribution of labels that follows: 57 negative, 62 neutral and 73 positive valence; and 13 low, 70 neutral and 109 high arousal. Files baseline_speech.ipynb and baseline_text.ipynb contain the code used to create the baseline emotion recognition models based on speech and text, respectively. The actual trained models for valence and arousal prediction are provided in folders models_speech and models_text.

Audio files in “Lecture” and “Emotions” are only provided to the users that complete the agreement file in section Usage Notes. Audio files are in Ogg Vorbis format at 16-bit and 44.1 kHz or 48 kHz. The total size of the “Audios” folder is about 213 MB.

Usage Notes

All the data included in the EMOVOME database is publicly available under the Creative Commons Attribution 4.0 International license. The only exception is the original raw audio files, for which an additional step is required as a security measure to safeguard the speakers' privacy. To request access, interested authors should first complete and sign the agreement file EMOVOME_agreement.pdf and send it to the corresponding author (jamarmo@htech.upv.es). The data included in the EMOVOME database is expected to be used for research purposes only. Therefore, the agreement file states that the authors are not allowed to share the data with profit-making companies or organisations. They are also not expected to distribute the data to other research institutions; instead, they are suggested to kindly refer interested colleagues to the corresponding author of this article. By agreeing to the terms of the agreement, the authors also commit to refraining from publishing the audio content on the media (such as television and radio), in scientific journals (or any other publications), as well as on other platforms on the internet. The agreement must bear the signature of the legally authorised representative of the research institution (e.g., head of laboratory/department). Once the signed agreement is received and validated, the corresponding author will deliver the "Audios" folder containing the audio files through a download procedure. A direct connection between the EMOVOME authors and the applicants guarantees that updates regarding additional materials included in the database can be received by all EMOVOME users.

Understanding the role of predators in food webs can be challenging in highly diverse predator/prey systems composed of small cryptic species. DNA based dietary analysis can supplement predator removal experiments and provide high resolution for prey identification. Here we use a metabarcoding approach to provide initial insights into the diet and functional role of coral-dwelling predatory fish feeding on small invertebrates. Fish were collected in Moorea (French Polynesia) where the BIOCODE project has generated DNA barcodes for numerous coral associated invertebrate species. Pyrosequencing data revealed a total of 292 Operational Taxonomic Units (OTU) in the gut contents of the arc-eye hawkfish (Paracirrhites arcatus), the flame hawkfish (Neocirrhites armatus) and the coral croucher (Caracanthus maculatus). One hundred forty-nine (51%) of them had species-level matches in reference libraries (>98% similarity) while 76 additional OTUs (26%) could be identified to higher taxonomic levels. Decapods that have a mutualistic relationship with Pocillopora and are typically dominant among coral branches, represent a minor contribution of the predators' diets. Instead, predators mainly consumed transient species including pelagic taxa such as copepods, chaetognaths and siphonophores suggesting non random feeding behavior. We also identified prey species known to have direct negative interactions with stony corals, such as Hapalocarcinus sp, a gall crab considered a coral parasite, as well as species of vermetid snails known for their deleterious effects on coral growth. Pocillopora DNA accounted for 20.8% and 20.1% of total number of sequences in the guts of the flame hawkfish and coral croucher but it was not detected in the guts of the arc-eye hawkfish. Comparison of diets among the three fishes demonstrates remarkable partitioning with nearly 80% of prey items consumed by only one predator. Overall, the taxonomic resolution provided by the metabarcoding approach highlights a highly complex interaction web and demonstrates that levels of trophic partitioning among coral reef fishes have likely been underestimated. Therefore, we strongly encourage further empirical approaches to dietary studies prior to making assumptions of trophic equivalency in food web reconstruction.

We combined the COI sequence data with legacy multigene sequence data to create a new, taxon-rich phylogeny for the Amaurobioidinae. We used sequences for four loci that have been used in previous studies on the subfamily: two mitochondrial loci, COI (658bp) and ribosomal subunit 16S (16S, 410bp); and two nuclear loci, Histone H3 (H3, 327bp) and ribosomal subunit 28S (28S, 839bp). We complemented the Amaurobioidinae data with sequences from several non-amaurobioidine anyphaenids and two clubionids as outgroups. Sequence alignment was performed using the MAFFT (ver. 7.308) plugin in Geneious, allowing MAFFT to automatically select an appropriate alignment strategy based on the properties of each locus, or with the online MAFFT server (https://mafft.cbrc.jp), which consistently selected the L-INS-i algorithm. Finally, alignments of the four loci were concatenated to construct a 2234 bp multigene sequence matrix containing 692 taxa, with about 55% missing/gap data (“full” matrix henceforth). To ensure that excessive missing data did not affect the resulting topology, we also constructed a reduced matrix by removing additional COI-only specimens so that each species and morphotype was represented by just one or two specimens for which all loci were available (where possible). After realignment, this reduced matrix was 2235 bp long, included 167 taxa, and had about 22% missing/gap data (“reduced” matrix henceforth). Phylogenetic analyses under maximum likelihood, including model selection, were then conducted with IQ-TREE 2. We performed phylogenetic analyses on both concatenated matrices (the full matrix and the reduced matrix) and on each individual locus. For model selection, we provided an initial scheme that partitioned the matrix by locus, and further partitioned the protein-coding loci (COI and H3) by codon position. We used ModelFinder and searched for the best partition scheme, all in IQ-TREE. The best models (partitions) for the full dataset were: GTR+F+I+G4 (16S), GTR+F+I+I+R4 (28S), TVM+F+I+I+R2 (COI-1), TIM2+F+R4 (COI-2), GTR+F+R5 (COI-3), TVMe+G4 (H3-1-H3-2), SYM+G4 (H3-3); and for the reduced dataset: GTR+F+I+G4 (16S), GTR+F+I+G4: (28S), GTR+F+I+G4: (COI-2), GTR+F+I+G4: (COI-3), TVM+F+I+G4: (COI-1, H3-2), GTR+F+I+G4: (H3-1), GTR+F+I+G4: (H3-3). For each dataset, once the best models and partitions were defined, we executed 10 independent replicates of tree calculations followed by 1000 ultrafast bootstrap replicates, and the replicate reaching the maximum likelihood was chosen. Phylogenetic analyses under parsimony were made with TNT, under equal weights, using the “new technology” search with default values, asking for 10 independent hits to the minimal length, and submitting the resulting trees to a round of TBR branch swapping.

Integrating Internet of Things (IoT) technology inside the cold supply chain can enhance transparency, efficiency, and quality, optimize operating procedures, and increase productivity. The integration of the IoT in this complicated setting is hindered by specific barriers that require thorough examination. Prominent barriers to IoT implementation in a cold supply chain, which is the main objective, are identified using a two–stage model. After reviewing the available literature on IoT implementation, 13 barriers were identified. The survey data were cross–validated for quality, and Cronbach’s alpha test was employed to ensure validity. This study applies the interpretative structural modeling technique in the first phase to identify the main barriers. Among these barriers, “regulatory compliance” and “cold chain networks” are the key drivers of IoT adoption strategies. MICMAC’s driving and dependence power element categorization helps evaluate barrier interactions. In the second phase of this study, a decision-making trial and evaluation laboratory methodology was employed to identify causal relationships between barriers and evaluate them according to their relative importance. Each cause is a potential drive, and if its efficiency can be enhanced, the system benefits as a whole. The findings provide industry stakeholders, governments, and organizations with significant drivers of IoT adoption to overcome these barriers and optimize the utilization of IoT technology to improve the effectiveness and reliability of the cold supply chain.