Multiple Templates Research Articles

MTGWNN: A Multi‐Template Graph Wavelet Neural Network Identification Model for Autism Spectrum Disorder

ABSTRACTFunctional magnetic resonance imaging (fMRI) has been widely applied in studying various brain disorders. However, current studies typically model regions of interest (ROIs) in brains with a single template. This approach generally examines only the connectivity between ROIs to identify autism spectrum disorder (ASD), ignoring the structural features of the brain. This study proposes a multi‐template graph wavelet neural network (GWNN) identification model for ASD called MTGWNN. First, the brain is segmented with multiple templates and the BOLD time series are extracted from fMRI data to construct brain networks. Next, a graph attention network (GAT) is applied to automatically learn interactions between nodes, capturing local information in the node features. These features are then further processed by a convolutional neural network (CNN) to learn global connectivity representations and achieve feature dimensionality reduction. Finally, the features and phenotypic data from each subject are integrated by GWNN to identify ASD at the optimal scale. Experimental results indicate that MTGWNN outperforms the comparative models. Testing on the public dataset ABIDE‐I achieved an accuracy (ACC) of 87.25% and an area under the curve (AUC) of 92.49%. MTGWNN effectively integrates brain network features from multiple templates, providing a more comprehensive characterization of brain abnormalities in patients with ASD. It incorporates population information from phenotypic data, which helps to compensate for the limited sample size of individual patients and improves the robustness and generalization of ASD identification.

AwareTrack: Object awareness for visual tracking via templates interaction

Current popular trackers, whether based on the Siamese network or Transformer, have focused their main work on relation modeling between the template and the search area, and on the design of the tracking head, neglecting the fundamental element of tracking, the template. Templates are often mixed with too much background information, which can interfere with the extraction of template features. To address the above issue, a template object-aware tracker (AwareTrack) is proposed. Through the information interaction between multiple templates, the attention of the templates can be truly focused on the object itself, and the background interference can be suppressed. To ensure that the foreground objects of the templates have the same appearance to the greatest extent, the concept of awareness templates is proposed, which consists of two close frames. In addition, an awareness templates sampling method based on similarity discrimination via Siamese network is also proposed, which adaptively determines the interval between two awareness templates, ensure the maximization of background differences in the awareness templates. Meanwhile, online updates to the awareness templates ensure that our tracker has access to the most recent features of the foreground object. Our AwareTrack achieves state-of-the-art performance on multiple benchmarks, particularly on the one-shot tracking benchmark GOT-10k, achieving the AO of 78.1%, which is a 4.4% improvement over OSTrack-384.

Prey detection by a stepwise visual template matching mechanism.

Predators can improve prey capture using a search image, and recent prey provide a visual template with which subsequent prey are compared. Considering trout feeding responses to mayfly prey of different sizes and phenological availability across years, we tested if changing relative abundances (ratios) of prey of the same species, but different body sizes, shifted trout feeding behaviour. For example, we hypothesized that a feeding switch from larger to smaller prey required continuous exposure to the novel smaller prey. The hypothesis that continuous exposure to novel small prey results in their acceptance was not supported. Rather, we discovered that trout identify novel prey using a dynamic stepwise visual neural template prey matching process, which involves the formation of focal prey template based on size or type, rejection of novel prey that do not match the size or type templates and modification of the existing or development of multiple prey templates that eventually enabled recognition of novel, small prey. We also discovered trout store multiple visual prey templates in memory. These results have implications for predator and prey dynamics, optimal foraging, the persistence of rare prey, prey species coexistence and predator selection on prey phenology.

Epitaxial Growth of Silver Clusters from Ag57 to Ag72 via Laminating Multiple Different Anion Templates.

The established capability of anion templates in precisely manipulating the size, geometry, and function of metal clusters is well acknowledged. However, the development of a systematic methodology for orchestrating the assembly of silver clusters, particularly those encompassing multiple distinct types of anion templates, remains elusive due to the formidable synthetic challenge. In this work, we report two novel silver clusters, Ag57 and Ag72, using two and three different anion templates, respectively. Ag57 features a gyroscope-like monovalent cation with an Ag3 triangle core sandwiched by one [SiW9O34]10- and a triad of Cl- anion templates. By intentionally introducing the third anion template, SO42-, the structure is expanded to the unprecedented Ag72 (with 15 silver atoms epitaxially grown on top of Ag57) resembling a tumbler, inside of which two Ag3 layers are laminated by one [SiW9O34]10-, seven Cl- and one SO42- anion templates in parallel with respect to longitudinal orientation. It is noteworthy that Ag72 exhibits remarkable structural complexity and represents a pioneering achievement as the first silver cluster incorporating three distinct types of anion templates. In addition, Ag72 demonstrates a significant advantage over Ag57, particular in terms of applications such as luminescent thermometers and remote laser ignition.

Epitaxial Growth of Silver Clusters from Ag57 to Ag72 via Laminating Multiple Different Anion Templates

The established capability of anion templates in precisely manipulating the size, geometry, and function of metal clusters is well acknowledged. However, the development of a systematic methodology for orchestrating the assembly of silver clusters, particularly those encompassing multiple distinct types of anion templates, remains elusive due to the formidable synthetic challenge. In this work, we report two novel silver clusters, Ag57 and Ag72, using two and three different anion templates, respectively. Ag57 features a gyroscope‐like monovalent cation with an Ag3 triangle core sandwiched by one [SiW9O34]10− and a triad of Cl− anion templates. By intentionally introducing the third anion template, SO42−, the structure is expanded to the unprecedented Ag72 (with 15 silver atoms epitaxially grown on top of Ag57) resembling a tumbler, inside of which two Ag3 layers are laminated by one [SiW9O34]10−, seven Cl− and one SO42− anion templates in parallel with respect to longitudinal orientation. It is noteworthy that Ag72 exhibits remarkable structural complexity and represents a pioneering achievement as the first silver cluster incorporating three distinct types of anion templates. In addition, Ag72 demonstrates a significant advantage over Ag57, particular in terms of applications such as luminescent thermometers and remote laser ignition.

Attentional templates for target features versus locations

Visual search is guided by visual working memory representations (i.e., attentional templates) that are activated prior to search and contain target-defining features (e.g., color). In the present study, we tested whether attentional templates can also contain spatial target properties (knowing where to look for) and whether attentional selection guided by such feature-specific templates is equally efficient than selection that is based on feature-specific templates (knowing what to look for). In every trial, search displays were either preceded by semantic color or location cues, indicating the upcoming target color or location, respectively. Qualitative differences between feature- and location-based template guidance were substantiated in terms of selection efficiency in low-load (one target color/location) versus high-load trials (two target colors/locations). Behavioral and electrophysiological (N2pc) measures of target selection speed and accuracy were combined for converging evidence. In line with previous studies, we found that color search was highly efficient, even under high-low conditions, when multiple attentional templates were activated to guide attentional selection in a spatially global fashion. Importantly, results in the location task almost perfectly mirrored the findings of the color task, suggesting that multiple templates for different target locations were activated concurrently when two possible target locations were task relevant. Our findings align with accounts that assume a common neuronal network during preparation for location and color search, but regard spatial and feature-based selection mechanisms as independent.

Temporal dynamics of multiple attentional template activation during preparation for search

Temporal dynamics of multiple attentional template activation during preparation for search

Development of an all-in-one real-time PCR assay for simultaneous detection of spotted fever group rickettsiae, severe fever with thrombocytopenia syndrome virus and hantaan virus prevalent in central China.

Central China has been reported to be one of the most important endemic areas of zoonotic infection by spotted fever group rickettsiae (SFGR), severe fever with thrombocytopenia syndrome virus (SFTSV) and hantaan virus (HTNV). Due to similar clinical symptoms, it is challenging to make a definite diagnosis rapidly and accurately in the absence of microbiological tests. In the present study, an all-in-one real-time PCR assay was developed for the simultaneous detection of nucleic acids from SFGR, SFTSV and HTNV. Three linear standard curves for determining SFGR-ompA, SFTSV-L and HTNV-L were obtained within the range of 101-106 copies/μL, with the PCR amplification efficiencies ranging from 93.46% to 96.88% and the regression coefficients R2 of >0.99. The detection limit was 1.108 copies/μL for SFGR-ompA, 1.075 copies/μL for SFTSV-L and 1.006 copies/μL for HTNV-L, respectively. Both the within-run and within-laboratory coefficients of variation on the cycle threshold (Ct) values were within the range of 0.53%-2.15%. It was also found there was no statistical difference in the Ct values between single template and multiple templates (PSFGR-ompA = 0.186, PSFTSV-L = 0.612, PHTNV-L = 0.298). The sensitivity, specificity, positive and negative predictive value were all 100% for determining SFGR-ompA and SFTSV-L, 97%, 100%, 100% and 99.6% for HTNV-L, respectively. Therefore, the all-in-one real-time PCR assay appears to be a reliable, sensitive, rapid, high-throughput and low cost-effective method to diagnose the zoonotic infection by SFGR, SFTSV and HTNV.

Comparing Gold-Standard Sanger Sequencing with Two Next-Generation Sequencing Platforms of HIV-1 gp160 Single Genome Amplicons.

Our goal was to assess the accuracy of next generation sequencing (NGS) compared with Sanger. We performed single genome amplification (SGA) of HIV-1 gp160 on extracted tissue DNA from two HIV+ individuals. Amplicons (n = 30) were sequenced with Sanger or reamplified with barcoded primers and pooled before sequencing using Oxford Nanopore Technologies (ONT) and Pacific Biosciences (PB). For each amplicon, a consensus sequence for NGS reads was obtained by (1) mapping reads to the Sanger sequence when available ("reference-based") or (2) mapping reads to a "pseudo-reference" sequence, i.e., a consensus sequence of a subset of NGS reads ("reference-free"). PB reads were clustered based on genetic similarity. A Sanger consensus sequence was obtained for 23/30 amplicons, for which all NGS consensus sequences were identical (n = 9) or nearly identical (n = 14) compared with Sanger. For the nine mismatches between Sanger/NGS, the nucleotide in the NGS sequence matched all other sequences from that patient. Of the 7/30 amplicons without a Sanger sequence, NGS sequences had ≥35 ambiguous calls in five amplicons and 0 ambiguitiesin two amplicons. Analysis of the electropherograms showed failure of a single sequencing primer for the latter two amplicons (consistent with a single template) and overlapping peaks for the other five (consistent with multiple templates). Clustering results closely followed the Sanger/NGS consensus results, where amplicons derived from a single template also had a single cluster and vice versa (with one exception, which could be the result of barcode misidentification). Representative sequences from the clusters contained 2-13 differences compared with Sanger/NGS. In summary, we show that both ONT and PB can produce amplicon consensus sequences with similar or higher accuracy compared with Sanger and, importantly, without the need for a known reference sequence. Clustering could be useful in some circumstances to predict or confirm the presence of multiple starting templates.

An Efficient Integrated Strategy for Comprehensive Metabolite Profiling of Sakurasosaponin from Aegiceras corniculatum in Rats.

Sakurasosaponin, a primary bioactive saponin from Aegiceras corniculatum, shows potential as an anti-cancer agent. However, there is a lack of information on its in vivo metabolism. This study aims to profile the in vivo metabolites of sakurasosaponin in rat feces, urine, and plasma after oral administration. An efficient strategy using ultra-high-performance liquid chromatography/quadrupole time-of-flight mass spectrometry was developed, which combined metabolic prediction, multiple mass defects filtering, and highresolution extracted ion chromatograms for rapid and systematic analysis. Firstly, a theoretical list of metabolites for sakurasosaponin was developed. This was done by considering the metabolic pathways of saponins. Next, the multiple mass defects filtering method was employed to identify potential metabolites in feces and urine, using the unique metabolites of sakurasosaponin as multiple mass defects filtering templates. Subsequently, a high-resolution extracted ion chromatogram was used to quickly determine the metabolites in rat plasma post-identification in feces and urine. Lastly, the analysis of accurate mass, typical neutral loss, and diagnostic ion of the candidate metabolites was carried out to confirm their structural elucidation, and metabolic pathways of sakurasosaponin in vivo were also proposed. In total, 30 metabolites were provisionally identified in feces, urine, and plasma. Analysis of metabolic pathways revealed isomerization, deglycosylation, oxidation, hydroxylation, sulfate conjugation, glucuronide conjugation, and other related reactions as the primary biotransformation reactions of sakurasosaponin in vivo. The findings demonstrate that the designed research strategy effectively minimizes matrix interference, prevents the omission of low-concentration metabolites, and serves as a foundation for the discovery of active metabolites of sakurasosaponin.

MixFormer: End-to-End Tracking With Iterative Mixed Attention.

Visual object tracking often employs a multi-stage pipeline of feature extraction, target information integration, and bounding box estimation. To simplify this pipeline and unify the process of feature extraction and target information integration, in this paper, we present a compact tracking framework, termed as MixFormer, built upon transformers. Our core design is to utilize the flexibility of attention operations, and we propose a Mixed Attention Module (MAM) for simultaneous feature extraction and target information integration. This synchronous modeling scheme allows us to extract target-specific discriminative features and perform extensive communication between target and search area. Based on MAM, we build our MixFormer trackers simply by stacking multiple MAMs and placing a localization head on top. Specifically, we instantiate two types of MixFormer trackers, a hierarchical tracker MixCvT, and a non-hierarchical simple tracker MixViT. For these two trackers, we investigate a series of pre-training methods and uncover the different behaviors between supervised pre-training and self-supervised pre-training in our MixFormer trackers. We also extend the masked autoencoder pre-training to our MixFormer trackers and design the new competitive TrackMAE pre-training technique. Finally, to handle multiple target templates during online tracking, we devise an asymmetric attention scheme in MAM to reduce computational cost, and propose an effective score prediction module to select high-quality templates. Our MixFormer trackers set a new state-of-the-art performance on seven tracking benchmarks, including LaSOT, TrackingNet, VOT2020, GOT-10 k, OTB100, TOTB and UAV123. In particular, our MixViT-L achieves AUC scores of 73.3% on LaSOT, 86.1% on TrackingNet and 82.8% on TOTB.

Rest-fMRI—A Potential Substitute for Task-fMRI?

Abstract Objective The aim of this study was to assess the reliability of resting-state functional magnetic resonance imaging (rest-fMRI) in mapping language areas for preoperative planning, versus standard task-based techniques, which are at times difficult to perform in clinical settings. Our study also aimed to evaluate the overlap between language areas identified through rest-fMRI and the standard task-fMRI, in neurosurgical cases. Materials and Methods Using a seed-based analysis of rest-fMRI with multiple template seeds, we identified functionally connected language regions in patients undergoing preoperative language mapping. Four language task paradigms (word, verb, picture, and semantics) were evaluated. We quantified the degree of overlap between language areas identified on rest-fMRI and task-fMRI, categorizing the results as more than 50% or less than 50% overlap. Results Seventy-seven percent of patients demonstrated an overlap exceeding 50% between rest- and task-fMRI maps, with the left Broca's area being the most frequently observed region of overlap. This finding was noted even in cases with lesions in Broca's or Wernicke's areas, highlighting the method's robustness. The verb task showed the best blood-oxygen-level dependent activity and overlap with rest-fMRI, highlighting its reliability. To identify a specific language area, the contralateral seed of the same area most commonly displayed connectivity with the area of interest. Conclusion Our findings demonstrate the potential of using rest-fMRI in accurately mapping eloquent language areas, in clinical settings The strong concordance observed, especially in the left Broca's area, underscores the reliability of this method. Further research and larger studies are essential to validate these results, potentially establishing the use of routine rest-fMRI, in clinical preoperative workup.

Tailoring porous structure in non-ionic polymer membranes using multiple templates for low-cost iron-lead single-flow batteries

Porous ion-selective membranes are promising alternatives for the expensive perfluorosulfonic acid membranes in redox flow batteries. In this work, novel non-ionic porous polyvinylidene fluoride-hexafluoro propylene membranes are designed for iron-lead single-flow batteries. The membranes are prepared using a multiple template approach, involving simultaneously using polyethylene glycol and dibutyl phthalate (DBP) as pore-forming templates. Their porous structure is finely tuned by adjusting the ratio of the two templates. As a result, dual-porous membranes bearing both macro and micropores are obtained. The H3520 membrane with modified porous structure attains a high proton conductivity of 43.5 mS·cm-1 and a relatively low ferric ion diffusion constant (8.61 × 10-8 cm2·min-1) and demonstrates the best balance between these performance-determining parameters (selectivity 5.04 × 105 S·min·cm-3, higher than that of the N115 membrane). Besides, performance tests of the iron-lead single-flow single cells equipped with the dual-porous membranes show a high energy efficiency, exceeding 87.2% at its rated current density, and outstanding cycling stability over 200 charge-discharge cycles. Altogether, the mixed template method presents a promising strategy to prepare high-performance and low-cost non-ionic membranes for redox flow batteries.

Activated long-term memory and visual working memory during hybrid visual search: Effects on target memory search and distractor memory.

In hybrid visual search, observers must maintain multiple target templates and subsequently search for any one of those targets. If the number of potential target templates exceeds visual working memory (VWM) capacity, then the target templates are assumed to be maintained in activated long-term memory (aLTM). Observers must search the array for potential targets (visual search), as well as search through memory (target memory search). Increasing the target memory set size reduces accuracy, increases search response times (RT), and increases dwell time on distractors. However, the extent of observers' memory for distractors during hybrid search is largely unknown. In the current study, the impact of hybrid search on target memory search (measured by dwell time on distractors, false alarms, and misses) and distractor memory (measured by distractor revisits and recognition memory of recently viewed distractors) was measured. Specifically, we aimed to better understand how changes in behavior during hybrid search impacts distractor memory. Increased target memory set size led to an increase in search RTs, distractor dwell times, false alarms, and target identification misses. Increasing target memory set size increased revisits to distractors, suggesting impaired distractor location memory, but had no effect on a two-alternative forced-choice (2AFC) distractor recognition memory test presented during the search trial. The results from the current study suggest a lack of interference between memory stores maintaining target template representations (aLTM) and distractor information (VWM). Loading aLTM with more target templates does not impact VWM for distracting information.

Robust 3D Tracking with Quality-Aware Shape Completion

3D single object tracking remains a challenging problem due to the sparsity and incompleteness of the point clouds. Existing algorithms attempt to address the challenges in two strategies. The first strategy is to learn dense geometric features based on the captured sparse point cloud. Nevertheless, it is quite a formidable task since the learned dense geometric features are with high uncertainty for depicting the shape of the target object. The other strategy is to aggregate the sparse geometric features of multiple templates to enrich the shape information, which is a routine solution in 2D tracking. However, aggregating the coarse shape representations can hardly yield a precise shape representation. Different from 2D pixels, 3D points of different frames can be directly fused by coordinate transform, i.e., shape completion. Considering that, we propose to construct a synthetic target representation composed of dense and complete point clouds depicting the target shape precisely by shape completion for robust 3D tracking. Specifically, we design a voxelized 3D tracking framework with shape completion, in which we propose a quality-aware shape completion mechanism to alleviate the adverse effect of noisy historical predictions. It enables us to effectively construct and leverage the synthetic target representation. Besides, we also develop a voxelized relation modeling module and box refinement module to improve tracking performance. Favorable performance against state-of-the-art algorithms on three benchmarks demonstrates the effectiveness and generalization ability of our method.

Beta-informativeness-diffusion multilayer graph embedding for brain network analysis.

Brain network analysis provides essential insights into the diagnosis of brain disease. Integrating multiple neuroimaging modalities has been demonstrated to be more effective than using a single modality for brain network analysis. However, a majority of existing brain network analysis methods based on multiple modalities often overlook both complementary information and unique characteristics from various modalities. To tackle this issue, we propose the Beta-Informativeness-Diffusion Multilayer Graph Embedding (BID-MGE) method. The proposed method seamlessly integrates structural connectivity (SC) and functional connectivity (FC) to learn more comprehensive information for diagnosing neuropsychiatric disorders. Specifically, a novel beta distribution mapping function (beta mapping) is utilized to increase vital information and weaken insignificant connections. The refined information helps the diffusion process concentrate on crucial brain regions to capture more discriminative features. To maximize the preservation of the unique characteristics of each modality, we design an optimal scale multilayer brain network, the inter-layer connections of which depend on node informativeness. Then, a multilayer informativeness diffusion is proposed to capture complementary information and unique characteristics from various modalities and generate node representations by incorporating the features of each node with those of their connected nodes. Finally, the node representations are reconfigured using principal component analysis (PCA), and cosine distances are calculated with reference to multiple templates for statistical analysis and classification. We implement the proposed method for brain network analysis of neuropsychiatric disorders. The results indicate that our method effectively identifies crucial brain regions associated with diseases, providing valuable insights into the pathology of the disease, and surpasses other advanced methods in classification performance.

A Method for Enhancing the Simulation Continuity of the Snesim Algorithm in 2D Using Multiple Search Trees

Multiple-point geostatistics (MPS) has more advantages than two-point geostatistics in reproducing the continuity of geobodies in subsurface reservoir modeling. For fluvial reservoir modeling, the more continuous a channel, the more consistent it is with geological knowledge in general, and fluvial continuity is also of paramount importance when simulating fluid flow. Based on the pixel-based MPS algorithm Snesim, this study proposes a method that utilizes multiple search trees (MSTs) to enhance simulation continuity in 2D fluvial reservoir modeling. The objective of the MST method is to capture complete data events from a training image (TI), which aims to achieve enhanced continuity in fluvial reservoir sublayer modeling. By resorting to search neighborhoods based on their proximity to the central node of the data template, multiple data templates that correspond to the MSTs will be generated. Here, four data templates were generated by arranging the relative search neighborhood coordinates in ascending and descending order with respect to the central node. Parallel computing was tried for the construction of the search trees. This work calculated the conditional probability distribution function (CPDF) of the simulating nodes by averaging the CPDFs derived from the MSTs, and double retrieval was employed to filter out the search trees that possessed an inaccurate local CPDF for the simulating nodes. In addition, the connected component labeling (CCL) method was introduced to evaluate the simulation continuity in MPS. The results indicated that the MST method can enhance the simulation continuity of the Snesim algorithm by reproducing the fine connectivity of channel facies in 2D fluvial reservoir modeling.

Prompt for extraction: Multiple templates choice model for event extraction

Event Extraction (EE) is an essential task in natural language processing that aims to mine events occurring in event mentions represent events using event records, which usually consist of event types, trigger words, argument elements corresponding to roles in the event types. Recently, prompt-based generative models have been developed to extract events. However, these prompt-based generative studies have ignored the fact that the strong language comprehension capability of the pre-trained language model (PLM) can analyze extract the potential role relationships in multiple templates for more information that can help extract argument elements. To determine the extended templates that can help the model for event extraction, we propose the multiple template choice model (MTCM), which designs an extended event type mining module to automatically mine the extended event types in the event mention uses the templates corresponding to the extended event types to interact with the template, corresponding to the currently to-be-extracted event type of event mention, in a multi-template information interaction, which gives the model more information guides the PLM for event extraction. To validate our model, we used two widely used datasets in the event extraction domain, ACE2005-EN ERE-EN. The experimental results show that our model achieves state-of-the-art performance on the ACE2005-EN dataset significantly improves the ERE dataset. In addition, according to the results, our model can be effectively adapted to low-resource environments.

Deep Fusion of Multi-Template Using Spatio-Temporal Weighted Multi-Hypergraph Convolutional Networks for Brain Disease Analysis.

Conventional functional connectivity network (FCN) based on resting-state fMRI (rs-fMRI) can only reflect the relationship between pairwise brain regions. Thus, the hyper-connectivity network (HCN) has been widely used to reveal high-order interactions among multiple brain regions. However, existing HCN models are essentially spatial HCN, which reflect the spatial relevance of multiple brain regions, but ignore the temporal correlation among multiple time points. Furthermore, the majority of HCN construction and learning frameworks are limited to using a single template, while the multi-template carries richer information. To address these issues, we first employ multiple templates to parcellate the rs-fMRI into different brain regions. Then, based on the multi-template data, we propose a spatio-temporal weighted HCN (STW-HCN) to capture more comprehensive high-order temporal and spatial properties of brain activity. Next, a novel deep fusion model of multi-template called spatio-temporal weighted multi-hypergraph convolutional network (STW-MHGCN) is proposed to fuse the STW-HCN of multiple templates, which extracts the deep interrelation information between different templates. Finally, we evaluate our method on the ADNI-2 and ABIDE-I datasets for mild cognitive impairment (MCI) and autism spectrum disorder (ASD) analysis. Experimental results demonstrate that the proposed method is superior to the state-of-the-art approaches in MCI and ASD classification, and the abnormal spatio-temporal hyper-edges discovered by our method have significant significance for the brain abnormalities analysis of MCI and ASD.

Making Use of Averaging Methods in MODELLER for Protein Structure Prediction.

Recent advances in protein structure prediction, driven by AlphaFold 2 and machine learning, demonstrate proficiency in static structures but encounter challenges in capturing essential dynamic features crucial for understanding biological function. In this context, homology-based modeling emerges as a cost-effective and computationally efficient alternative. The MODELLER (version 10.5, accessed on 30 November 2023) algorithm can be harnessed for this purpose since it computes intermediate models during simulated annealing, enabling the exploration of attainable configurational states and energies while minimizing its objective function. There have been a few attempts to date to improve the models generated by its algorithm, and in particular, there is no literature regarding the implementation of an averaging procedure involving the intermediate models in the MODELLER algorithm. In this study, we examined MODELLER's output using 225 target-template pairs, extracting the best representatives of intermediate models. Applying an averaging procedure to the selected intermediate structures based on statistical potentials, we aimed to determine: (1) whether averaging improves the quality of structural models during the building phase; (2) if ranking by statistical potentials reliably selects the best models, leading to improved final model quality; (3) whether using a single template versus multiple templates affects the averaging approach; (4) whether the "ensemble" nature of the MODELLER building phase can be harnessed to capture low-energy conformations in holo structures modeling. Our findings indicate that while improvements typically fall short of a few decimal points in the model evaluation metric, a notable fraction of configurations exhibit slightly higher similarity to the native structure than MODELLER's proposed final model. The averaging-building procedure proves particularly beneficial in (1) regions of low sequence identity between the target and template(s), the most challenging aspect of homology modeling; (2) holo protein conformations generation, an area in which MODELLER and related tools usually fall short of the expected performance.