Size Selection Techniques Research Articles

Sample Size Considerations for Fine-Tuning Large Language Models for Named Entity Recognition Tasks: Methodological Study.

Large language models (LLMs) have the potential to support promising new applications in health informatics. However, practical data on sample size considerations for fine-tuning LLMs to perform specific tasks in biomedical and health policy contexts are lacking. This study aims to evaluate sample size and sample selection techniques for fine-tuning LLMs to support improved named entity recognition (NER) for a custom data set of conflicts of interest disclosure statements. A random sample of 200 disclosure statements was prepared for annotation. All "PERSON" and "ORG" entities were identified by each of the 2 raters, and once appropriate agreement was established, the annotators independently annotated an additional 290 disclosure statements. From the 490 annotated documents, 2500 stratified random samples in different size ranges were drawn. The 2500 training set subsamples were used to fine-tune a selection of language models across 2 model architectures (Bidirectional Encoder Representations from Transformers [BERT] and Generative Pre-trained Transformer [GPT]) for improved NER, and multiple regression was used to assess the relationship between sample size (sentences), entity density (entities per sentence [EPS]), and trained model performance (F1-score). Additionally, single-predictor threshold regression models were used to evaluate the possibility of diminishing marginal returns from increased sample size or entity density. Fine-tuned models ranged in topline NER performance from F1-score=0.79 to F1-score=0.96 across architectures. Two-predictor multiple linear regression models were statistically significant with multiple R2 ranging from 0.6057 to 0.7896 (all P<.001). EPS and the number of sentences were significant predictors of F1-scores in all cases ( P<.001), except for the GPT-2_large model, where EPS was not a significant predictor (P=.184). Model thresholds indicate points of diminishing marginal return from increased training data set sample size measured by the number of sentences, with point estimates ranging from 439 sentences for RoBERTa_large to 527 sentences for GPT-2_large. Likewise, the threshold regression models indicate a diminishing marginal return for EPS with point estimates between 1.36 and 1.38. Relatively modest sample sizes can be used to fine-tune LLMs for NER tasks applied to biomedical text, and training data entity density should representatively approximate entity density in production data. Training data quality and a model architecture's intended use (text generation vs text processing or classification) may be as, or more, important as training data volume and model parameter size.

Sonication-assisted liquid exfoliation and size-dependent properties of magnetic two-dimensional α-RuCl3

Originating from the hexagonal arrangement of magnetic ions in the presence of strong spin orbit coupling, α-RuCl3 is considered as model system for the Kitaev-Heisenberg model. While the magnetic properties of α-RuCl3 have been studied in bulk single crystals or micromechanically-exfoliated nanosheets, little is known about the nanosheets’ properties after exfoliation by techniques suitable for mass production such as liquid phase exfoliation (LPE). Here, we demonstrate sonication-assisted LPE on α-RuCl3 single crystals in an inert atmosphere. Coupled with centrifugation-based size selection techniques, the accessible size- and thickness range is quantified by statistical atomic force microscopy. Individual nanosheets obtained after centrifugation-based size selection are subjected to transmission electron microscopy to confirm their structural integrity after the exfoliation. The results are combined with bulk characterisation methods, including Raman and x-ray photoelectron spectroscopy, and powder diffraction experiments to evaluate the structural integrity of the nanosheets. We report changes of the magnetic properties of the nanomaterial with nanosheet size, as well as photospectroscopic metrics for the material concentration and average layer number. Finally, a quantitative analysis on environmental effects on the nanomaterial integrity is performed based on time and temperature dependent absorbance spectroscopy revealing a relatively slow decay (half-life of ∼2000 h at 20 °C), albeit with low activation energies of 6–20 kJ mol−1.

A new adaptive window‐based guided filtering and interpolation for polarization image demosaicing

AbstractThis paper presents a new adaptive window‐based guided filtering and interpolation for polarization image demosaicing. Considering the correlation between the polarization channels, the authors use the total intensity S0 as a guide image to guide the interpolation of all channels. The guided filtering is performed with a new adaptive window size selection technique. A novel iterative residual interpolation is proposed to construct accurate full resolution polarization. Experiments are conducted with the sample images from the demosaicking algorithms dataset. Objective evaluations demonstrate that the proposed algorithm can reduce root mean squared error by at least 27% over other methods. Subjective visual inspection can also verify the improvements in the clarity of objects in angle and degree of the linear polarization images.

Loop interchange and tiling for multi-dimensional loops to minimize write operations on NVMs

Non-volatile memory (NVM) is expected to be the second tier of memory in two-tier memory systems. However, because of the limited write endurance, it is vital to reduce the number of writes on NVM. Large-scale nested loops are the performance bottleneck in programs since the data cannot be held on the first tier of memory and then causes many write operations on NVM. Loop tiling groups iterations and loop interchange changes the execution order of the loop to improve data locality and thus reduce communication with NVM. However, research that combines loop interchange and loop tiling for minimizing writes on NVM is uncommon. In this paper, we propose a new loop tiling scheme and combine the loop interchange to solve these issues. Specifically, we propose a strategy to generate the legal tile shape, which is a parallelogram, first. Then, we propose an optimal tile size selection technique to minimize the write operations on NVM. In addition, we adopt the loop interchange technique to help loop tiling generate an optimal tile size for multi-dimensional loops. Finally, we schedule the accessing operations and computations in a pipeline fashion to cover the NVM latency. Experiments show that the proposed scheme can reduce the write on NVM efficiently. In addition, for 2-dimensional loops, NVM latency can be completely hidden.

Does a Change in Device Design Alter Device Size Selection? A Comparison of Conventional and Occlutech Duct Occluder Designs.

The Occlutech duct occluder (ODO) with a novel design has been available for closure of patent ductus arteriosus (PDA) since 2011. Available data on initial experience with the ODO have shown that operators continue to choose device sizes based on their experience with conventional duct occluders (CDO). This study postulated that occlusion of the pulmonary arterial (PA) end of the ductus should achieve satisfactory PDA closure without additional complications. Accordingly, the size of the ODO was chosen such that the larger PA diameter in the new design exceeded the PA end of the ductus by 2-3 mm and 2-4 mm for normotensive and hypertensive ducts, respectively. It sought to examine the feasibility and safety of such an approach, and compared ODO and CDO devices with respect to device sizes deployed, to ascertain if the newer design had any advantages. This prospective study enrolled 105 infants, children and a few adults with various duct morphologies and PA pressures for ductal closure using the ODO from 2018-2020. The control group consisted of 105 ducts closed with CDO. A comparison of the two groups with respect to duct diameter and implanted device sizes was performed using appropriate statistical software. The study found that 40% of the ducts had moderate-to-severe pulmonary hypertension. Most ducts measured between 3-4 mm in both groups. The mean weight of the subjects in both groups was 12 kg. Although there was no change in the way that duct occluders were chosen in both groups (2-3 mm over the pulmonary artery diameter), ODO could be significantly downsized in most ducts measuring between 2.5-6.4 mm in diameter (p<0.05). This study found successful closure of relatively larger PDA in infants and children using smaller ODO compared with CDO, without embolisation or aortic narrowing. It also demonstrated that a different device size selection technique may be safely employed with the ODO.

Deriving Right Sample Size and Choosing an Appropriate Sampling Technique to Select Samples from the Research Population During Ph.D. Program in India

Purpose: The purpose of this article is to explain standard formulas available for deriving sample size, the essence of every component of formulas, and available techniques for selecting samples from the research population in turn, guiding the Ph.D. scholars to finalize appropriate sample size and sampling technique. Design/Methodology/Approach: Postmodernism philosophical paradigm; Inductive research approach; Observation data collection method; Longitudinal data collection time frame; Qualitative data analysis. Findings/Result: As long as the Ph.D. scholars can understand an appropriate sample size and available sampling techniques and make mindful choices of sample size and sampling technique across various stages/phases of the research to answer their research questions they will be able to determine (on their own) all the other choices in succeeding steps of doctoral-level research such as i) data collection instrument and iii) data analysis techniques. Originality/Value: There is a vast literature about how to derive the sample size and how to select samples from the research population. However, only a few have explained them together comprehensively which is conceivable to Ph.D. scholars. In this article, we have attempted to explain every component of sample size formulas and capture most of the sampling techniques briefly that would enable Ph.D. scholars in India to glance through and make a scholarly choice of appropriate sample size and sample selection techniques. Paper Type: Conceptual.

Size Selection and Size‐Dependent Optoelectronic and Electrochemical Properties of 2D Titanium Carbide (Ti3C2Tx) MXene

Abstract2D materials produced via liquid‐phase exfoliation commonly exhibit broad size and thickness distributions, where size selection is required to obtain nanosheets with desired dimensions and properties. Here, size selection of 2D Ti3C2Tx MXene is systematically performed using liquid cascade centrifugation (LCC) and sedimentation‐based separation (SBS) techniques. Based on statistical atomic force and electron microscopy analyses, it is found that the SBS technique offers stronger separation power compared to the LCC technique with an expense of lower yield. Moreover, predominantly monolayer MXene flakes are obtained after centrifugation at low speeds (≤ 3000 rpm) for 15 min for both techniques. Films fabricated from large MXene flakes with the average flake length of 4.2 µm show high electrical figure of merit of 7.8 and improved rate performance as a cathode in Zn‐ion capacitor. This work serves as a guideline for choosing size selection techniques and conditions for MXene processing to obtain flakes with certain sizes and desired properties.

Problems of the Grid Size Selection in Differential Box-Counting (DBC) Methods and an Improvement Strategy.

The differential box-counting (DBC) method is useful for determining the fractal dimension of grayscale images. It is simple to learn and implement and has been extensively utilized. However, this approach has several problems, such as over- or undercounting the number of boxes due to inappropriate parameter choices, limiting the calculation accuracy. Many studies have been conducted to increase the algorithm’s computational accuracy by improving the calculating parameters of the differential box-counting method. The grid size is a crucial parameter for the DBC method. Generally, there are two typical ways for selecting the grid size in relevant studies: consecutive integer and divisors of image size. However, both methods for grid size selection are problematic. The consecutive integer method cannot partition the image entirely and will result in the undercounting of boxes; the divisors of image size can partition the image completely. However, this method uses fewer grid sizes to compute fractal dimensions and has a relatively huge distance error (DE). To address the shortcomings of the above-mentioned two approaches, this research presents an improved grid size selection strategy. The improved method enhances computational accuracy by computing the discarded image edge areas in the consecutive integer method, allowing the original image information to be used as thoroughly as the divisor strategy. Based on fractional Brownian motion (FBM), Brodatz, and Aerials image sets, the accuracy of the three grid size selection techniques (consecutive integer method, divisors of image size method, and the improved algorithm) to compute the fractal dimension is then compared. The results reveal that, compared to the two prior techniques, the revised algorithm described in this study minimizes the distance error and increases the accuracy of the fractal dimension computation.

Medicolegal Sidebar: Are Implant Sales Reps in the Operating Room Legally Untouchable?

Medicolegal Sidebar: Are Implant Sales Reps in the Operating Room Legally Untouchable?

A novel low-cost method for generalized split inverse problem of finite family of demimetric mappings

The purpose of this paper is to introduce iterative method for solving generalized split inverse problem given as a task of finding a point which belongs to the intersection of finite family of fixed point sets of demimetric mappings such that its image under a finite number of linear transformations belongs to the intersection of another finite family of fixed point sets of demimetric mappings in the image space. The proposed algorithm is formulated in low-cost sequential computing method with step size selection technique. The strong convergence theorem of the proposed algorithm is derived under the appropriate assumptions. The result presented in the paper generalized several results in the literature. Numerical example is given to illustrate the efficiency and performance of the proposed iterative method.

Comprehensive review and assessment of computational methods for predicting RNA post-transcriptional modification sites from RNA sequences.

RNA post-transcriptional modifications play a crucial role in a myriad of biological processes and cellular functions. To date, more than 160 RNA modifications have been discovered; therefore, accurate identification of RNA-modification sites is fundamental for a better understanding of RNA-mediated biological functions and mechanisms. However, due to limitations in experimental methods, systematic identification of different types of RNA-modification sites remains a major challenge. Recently, more than 20 computational methods have been developed to identify RNA-modification sites in tandem with high-throughput experimental methods, with most of these capable of predicting only single types of RNA-modification sites. These methods show high diversity in their dataset size, data quality, core algorithms, features extracted and feature selection techniques and evaluation strategies. Therefore, there is an urgent need to revisit these methods and summarize their methodologies, in order to improve and further develop computational techniques to identify and characterize RNA-modification sites from the large amounts of sequence data. With this goal in mind, first, we provide a comprehensive survey on a large collection of 27 state-of-the-art approaches for predicting N1-methyladenosine and N6-methyladenosine sites. We cover a variety of important aspects that are crucial for the development of successful predictors, including the dataset quality, operating algorithms, sequence and genomic features, feature selection, model performance evaluation and software utility. In addition, we also provide our thoughts on potential strategies to improve the model performance. Second, we propose a computational approach called DeepPromise based on deep learning techniques for simultaneous prediction of N1-methyladenosine and N6-methyladenosine. To extract the sequence context surrounding the modification sites, three feature encodings, including enhanced nucleic acid composition, one-hot encoding, and RNA embedding, were used as the input to seven consecutive layers of convolutional neural networks (CNNs), respectively. Moreover, DeepPromise further combined the prediction score of the CNN-based models and achieved around 43% higher area under receiver-operating curve (AUROC) for m1A site prediction and 2-6% higher AUROC for m6A site prediction, respectively, when compared with several existing state-of-the-art approaches on the independent test. In-depth analyses of characteristic sequence motifs identified from the convolution-layer filters indicated that nucleotide presentation at proximal positions surrounding the modification sites contributed most to the classification, whereas those at distal positions also affected classification but to different extents. To maximize user convenience, a web server was developed as an implementation of DeepPromise and made publicly available at http://DeepPromise.erc.monash.edu/, with the server accepting both RNA sequences and genomic sequences to allow prediction of two types of putative RNA-modification sites.

On the Barzilai‐Borwein basic scheme in FFT‐based computational homogenization

SummaryBuilding upon the equivalence of the basic scheme in the work of Moulinec and Suquet with gradient descent methods, we investigate the effect of using the celebrated Barzilai‐Borwein step size selection technique in this context. We provide an overview of recent convergence theory and present efficient implementations in the context of computational micromechanics, with and without globalization. In contrast to polarization schemes and fast gradient methods, no lower bound on the eigenvalues of the material tangent is necessary for the Barzilai‐Borwein scheme. We demonstrate the power of the proposed method for linear elastic and inelastic large scale problems with finite and infinite material contrast.

Tuning iteration space slicing based tiled multi-core code implementing Nussinov\u2019s RNA folding

BackgroundRNA folding is an ongoing compute-intensive task of bioinformatics. Parallelization and improving code locality for this kind of algorithms is one of the most relevant areas in computational biology. Fortunately, RNA secondary structure approaches, such as Nussinov’s recurrence, involve mathematical operations over affine control loops whose iteration space can be represented by the polyhedral model. This allows us to apply powerful polyhedral compilation techniques based on the transitive closure of dependence graphs to generate parallel tiled code implementing Nussinov’s RNA folding. Such techniques are within the iteration space slicing framework – the transitive dependences are applied to the statement instances of interest to produce valid tiles. The main problem at generating parallel tiled code is defining a proper tile size and tile dimension which impact parallelism degree and code locality.ResultsTo choose the best tile size and tile dimension, we first construct parallel parametric tiled code (parameters are variables defining tile size). With this purpose, we first generate two nonparametric tiled codes with different fixed tile sizes but with the same code structure and then derive a general affine model, which describes all integer factors available in expressions of those codes. Using this model and known integer factors present in the mentioned expressions (they define the left-hand side of the model), we find unknown integers in this model for each integer factor available in the same fixed tiled code position and replace in this code expressions, including integer factors, with those including parameters. Then we use this parallel parametric tiled code to implement the well-known tile size selection (TSS) technique, which allows us to discover in a given search space the best tile size and tile dimension maximizing target code performance.ConclusionsFor a given search space, the presented approach allows us to choose the best tile size and tile dimension in parallel tiled code implementing Nussinov’s RNA folding. Experimental results, received on modern Intel multi-core processors, demonstrate that this code outperforms known closely related implementations when the length of RNA strands is bigger than 2500.

(Invited) Liquid-Exfoliated Transition Metal Dichalcogenides: A Story of Excitons, Spectroscopic Metrics and Functionalisation

Liquid exfoliation has become an important production technique to give access to large quantities of two-dimensional nanosheets in colloidal dispersion.[1] Importantly, this is a highly versatile technique that can be applied to numerous layered materials beyond graphene such as transition metal dichalcogenides, III-VI semiconductors, black phosphorus, layered oxides to name just a few. These can be cast into films and composites and have proven useful in a number of application areas. However, as produced dispersions contain a nanosheets of various sizes and thicknesses making post-exfoliation size selection extremely important to tap the full potential in the future. We have previously developed a universal and efficient size selection technique termed liquid cascade centrifugation.[2] Optimisation of this centrifugation procedure allowed us to produce monolayer-rich dispersions with monolayer volume fractions up to 75%. This was made possible because we have realised that nanosheet size and thickness, as well as monolayer content can be quantitatively extracted from for example optical extinction spectra due to edge and confinement effects greatly facilitating and accelerating the characterisation.[2,3] We now use this basic understanding and these improvements in sample quality to systematically investigate the effect of the chemical environment on the optical properties of the liquid-exfoliated nanosheets. For example, we show that the A-exciton energy depends on the layer number in a similar way for a number of different TMDs. Furthermore, dielectric screening from additives and solvents can be tracked via this A-exciton layer number dependence making this an ideal fingerprint for noncovalent functionalisation.[4] Last but not least, we show that the ability to dramatically increase the surface area of the layered materials on exfoliation can be used to explore new chemistry and to controllably functionalise either basal plane [5,6] or edge sites. [1] Bonaccorso, F., Bartolotta, A.; Coleman, J.N., Backes, C., Adv. Mater. 2016, 28, 6136-6166. [2] Backes, C.; Szydłowska, B.M.; … Coleman, J. N., ACS Nano 2016, 10, 1589-1601. [3] Backes, C.; Smith, R. J.; … Coleman, J. N., Nature Communications 2014, 5, 4576. [4] Vega-Mayoral, V.; Backes, C.; …. Coleman, J. N., Adv. Func. Mater. 2016, 7, 1028-1039. [5] Backes, C.; Berner, N. C.; …. McDonald, A., Angew. Ch. Int. Ed.. 2015, 54, 1-6. [6] Knirsch, K.; Berner, N. C.; …. Backes, C., ACS Nano 2015, 9, 6018-6030.

Habitat type and ambient temperature contribute to bill morphology

Avian bills are iconic structures for the study of ecology and evolution, with hypotheses about the morphological structure of bills dating back to Darwin. Several ecological and physiological hypotheses have been developed to explain the evolution of the morphology of bill shape. Here, we test some of these hypotheses such as the role of habitat, ambient temperature, body size, intraspecific competition, and ecological release on the evolution of bill morphology. Bill morphology and tarsus length were measured from museum specimens of yellow warblers, and grouped by habitat type, sex, and subspecies. We calculated the mean maximum daily temperature for the month of July, the hottest month for breeding specimens at each collecting location. Analysis of covariance models predicted total bill surface area as a function of sex, habitat type, body size, and temperature, and model selection techniques were used to select the best model. Habitat, mangrove forests compared with inland habitats, and climate had the largest effects on bill size. Coastal wetland habitats and island populations of yellow warblers had similar bill morphology, both of which are larger than mainland inland populations. Temperate but not tropical subspecies exhibited sexual dimorphism in bill morphology. Overall, this study provides evidence that multiple environmental factors, such as temperature and habitat, contribute to the evolution of bill morphology.

Measuring the lateral size of liquid-exfoliated nanosheets with dynamic light scattering

We have developed an in situ method to estimate the lateral size of exfoliated nanosheets dispersed in a liquid. Using standard liquid exfoliation and size-selection techniques, we prepared a range of dispersions of graphene, MoS2 and WS2 nanosheets with different mean lateral sizes. The mean nanosheet length was measured using transmission electron microscopy (TEM) to vary from ∼40 nm to ∼1 μm. These dispersions were characterized using a standard dynamic light scattering (DLS) instrument. We found a well-defined correlation between the peak of the particle size distribution as outputted by the DLS instrument and the nanosheet length as measured by TEM. This correlation is consistent with the DLS instrument outputting the radius of a sphere with volume equal to the mean nanosheet volume. This correlation allows the mean nanosheet length to be extracted from DLS data.

A combined size sorting strategy for monodisperse plasmonic nanostructures

The fabrication of highly monodisperse silica coated Au NPs by the microemulsion approach and the selection of the nanostructure morphology have been described. Several experimental conditions, synthetic parameters and post-preparative strategies such as reaction time, precursor concentration, size selection techniques and NP surface treatments have been suitably investigated in order to fabricate Au and Au@SiO2 NPs with peculiar and tuneable plasmonic properties that strongly depend on the specific size distribution and nanostructure morphology. In particular, size selected precipitation of oleylamine-capped Au NPs by antisolvent titration has successfully offered a strategy to discriminate and collect monodisperse fractions with different average size and narrow size distribution. Moreover, for the first time, a deep insight into the microemulsion mechanism for the silica shell growth has been provided, highlighting the critical role played by the density of oleylamine at the Au NP surface. Specifically the capping agent has been demonstrated to strongly determine the multiplicity of the core in the final Au@SiO2 nanostructures. Density gradient centrifugation has been finally performed to sort the achieved Au@SiO2 NPs with different morphologies, which was ultimately able to recover a significant fraction formed of two Au NPs in one silica shell. A systematic characterization of the Au and Au@SiO2 NPs has been carried out by complementary morphological and spectroscopic techniques. These deeply investigated materials, with tuneable plasmonic properties, have been proposed as versatile building blocks useful for the design and fabrication of plasmonic and photonic structures as well as metamaterials for device applications.

Isolation of Rat Portal Fibroblasts by &lt;em&gt;In situ&lt;/em&gt; Liver Perfusion

Liver fibrosis is defined by the excessive deposition of extracellular matrix by activated myofibroblasts. There are multiple precursors of hepatic myofibroblasts, including hepatic stellate cells, portal fibroblasts and bone marrow derived fibroblasts 1. Hepatic stellate cells have been the best studied, but portal fibroblasts are increasingly recognized as important contributors to the myofibroblast pool, particularly in biliary fibrosis 2. Portal fibroblasts undergo proliferation in response to biliary epithelial injury, potentially playing a key role in the early stages of biliary scarring 3-5. A method of isolating portal fibroblasts would allow in vitro study of this cell population and lead to greater understanding of the role portal fibroblasts play in biliary fibrosis.Portal fibroblasts have been isolated using various techniques including outgrowth 6, 7 and liver perfusion with enzymatic digestion followed by size selection 8. The advantage of the digestion and size selection technique compared to the outgrowth technique is that cells can be studied without the necessity of passage in culture. Here, we describe a modified version of the original technique described by Kruglov and Dranoff 8 for isolation of portal fibroblasts from rat liver that results in a relatively pure population of primary cells.

Isolation of Rat Portal Fibroblasts by &lt;em&gt;In situ&lt;/em&gt; Liver Perfusion

Liver fibrosis is defined by the excessive deposition of extracellular matrix by activated myofibroblasts. There are multiple precursors of hepatic myofibroblasts, including hepatic stellate cells, portal fibroblasts and bone marrow derived fibroblasts 1. Hepatic stellate cells have been the best studied, but portal fibroblasts are increasingly recognized as important contributors to the myofibroblast pool, particularly in biliary fibrosis 2. Portal fibroblasts undergo proliferation in response to biliary epithelial injury, potentially playing a key role in the early stages of biliary scarring 3-5. A method of isolating portal fibroblasts would allow in vitro study of this cell population and lead to greater understanding of the role portal fibroblasts play in biliary fibrosis. Portal fibroblasts have been isolated using various techniques including outgrowth 6, 7 and liver perfusion with enzymatic digestion followed by size selection 8. The advantage of the digestion and size selection technique compared to the outgrowth technique is that cells can be studied without the necessity of passage in culture. Here, we describe a modified version of the original technique described by Kruglov and Dranoff 8 for isolation of portal fibroblasts from rat liver that results in a relatively pure population of primary cells.

Identification by cDNA Cloning of Abundant sRNAs in a Human-Avirulent Yersinia pestis Strain Grown Under Five Different Growth Conditions

sRNA regulation is supposedly involved in the stress response of a pathogen during infection. Yersinia pestis, the etiologic agent of plague, must encounter temperature and microenvironment changes, given its lifestyle. Here, we used the cDNA cloning approach to discover full-length sRNA candidates that are highly expressed in Y. pestis under five different growth conditions. The cDNA cloning approach was improved by combining the traditional cDNA library construction with the prevalent rapid amplification of cDNA ends and RNA size selection techniques. In total, 43 RNA species, including six previously annotated sRNAs, were identified. Of these, 25 sRNAs were encoded on the antisense strand of the annotated genes. Interestingly, two of these sRNAs were found on the complementary strand of noncoding RNAs. In addition, eight novel sRNAs encoded in the intergenic regions were also revealed. Ten sRNA candidates chosen for the northern blot analysis were successfully detected. Analysis of the expression patterns of 29 candidate sRNAs showed that 24 sRNAs are highly abundant in Y. pestis upon entry into the stationary growth phase. Our preliminary attempt at screening the novel sRNA candidates will lay the foundation for understanding the roles of sRNAs in Y. pestis physiology and pathogenesis.