Ranked List Research Articles

Resources Utilized by Otolaryngology Fellowship Directors When Determining Reputations of Residency Programs.

Residency program reputation is consistently reported as an important factor by fellowship directors when considering applicants. This study sets out to determine resources fellowship directors rely on when determining residency program reputation. Cross-sectional study. Using an anonymous online survey of all 2022 Otolaryngology (OHNS) fellowship program directors. The 13-question survey sought to assess fellowship director's perspectives and resource utilization when determining residency program reputation. Representing all OHNS fellowship program directors and co-directors (N = 287), 103 responded to our survey, response rate 35.9%. Most participants reported that residency reputation was important for fellowship candidacy. On a Likert scale of 1 to 5, 1 being most important and 5 being not important, personal knowledge of the residency program (2.03 out of 5) and program faculty/mentor reputation (2.09 out of 5) were the most important factors cited. 63% were unfamiliar with the survey methodology of Doximity Residency Navigator (DRN), while 53% contributed to DRN by filling out surveys. Nearly all fellowship directors (N = 100, 97%) reported their rank list was not influenced by DRN. Most fellowship directors reported that US News and World Report (USNWR) and DRN were neither consistent nor inconsistent with their perceptions of residency reputations (38% and 56%, respectively), suggesting ambivalence toward these resources. Residency reputation is important for fellowship directors when evaluating fellowship candidates. Directors do not rely on USNWR, National Institute of Health (NIH) ranking, or DRN when gauging residency reputation, but rather personal knowledge of the applicant's residency program or reputation of the otolaryngology faculty.

Introducing re-weighted range voting in clinical practice guideline prioritization: Development and testing of the re-weighted priority-setting (REPS) tool.

We aimed to develop and test a tool based on the re-weighted range voting mechanism to prioritize items (i.e. key questions) in a priority-setting assessment for clinical practice guidelines. The secondary aim was to provide methodological context of the tool. We iteratively developed the tool and used qualitative methods (i.e. think-aloud and semi-structured interviews) to test the tool's usability and make adjustments accordingly. An observational approach was used to test the tool's outcome satisfaction in a real-world priority-setting assessment within a rare-disease guideline of a European Reference Network and under four different conditions in the tool. Four guideline methodologists tested the usability of the tool. The real-world testing was performed with a guideline panel consisting of a core working group, five expertise working groups, and a working group with patient representatives. Thirty-one panel members assigned scores in the priority-setting assessment. Seventeen panel members rated the priority-setting outcome, and sixteen panel members rated the outputs generated under the four conditions. Upon initial use, guideline methodologists found the tool to be quite overwhelming. However, with some initial effort they were able to easily identify the tool's structure. Based on observations and feedback, the tool was further refined and user guidance was developed. Guideline panel members expressed (high) satisfaction with the priority-setting outcome. They particularly preferred the condition when using mean subgroup scores as input or employing aggressive penalties in the weighting method to determine the outputs. The tool generates a ranked list of items and offers flexibility for different choices in priority-setting assessments as long as its input format requirements are met. Although it is not a consensus method, the tool assists in narrowing down a set of priority items. Additional steps in the priority-setting assessment can lead to a consensus being reached regarding the final outcome.

Abstract LB193: Enhanced single cell DNA methylation analysis using combinatorial indexing

Abstract The epigenetic landscape of the human brain undergoes a plethora of changes during malignant transformation. In recent years, DNA methylation-based epigenetic modifications have been widely studied using traditional techniques like bisulfite sequencing and enzymatic methyl sequencing (EM-seq). However, these methods analyze bulk cell populations and lack the granularity of single-cell analysis. Although advances in single-cell analysis have revolutionized our understanding of cellular heterogeneity and functional diversity within complex biological systems, they are still costly, low throughput, and laborious. To address these challenges, ScaleBio has pioneered combinatorial indexing technology, enabling a significant increase in cell throughput. This method utilizes the cell itself as a compartment to perform 2-3 rounds of sequential barcoding in a plate-based workflow, eliminating the need for complex instrumentation. This technology has been successfully adapted to assess DNA methylation at the single-cell level offering a robust, affordable, high-throughput protocol that enhances yield, diversity, and coverage. In this study we used ScaleBio's single-cell methylation kit to investigate DNA methylation patterns during oncogenesis using cancer cells with widespread DNA methylation changes, such as human isocitrate dehydrogenase (IDH) mutant glioma cells at the single-cell level. The IDH gene family, comprising of IDH1, IDH2, and IDH3, encodes enzymes involved in the tricarboxylic acid (TCA) cycle. IDH1/2 mutations are present in over 80% of low-grade gliomas while IDH-mutant gliomas constitute about 1/5th of all adult diffuse gliomas, thus making them one of the most common brain tumor subtypes. As such, IDH mutations have emerged as attractive therapeutic targets for glioma treatment. By uncovering DNA methylation patterns at the single cell level, we provide here an epigenetic map to better understand this complex disease and better inform clinical discoveries. We achieved high cell recovery and robust cytosine coverage throughout our analysis of single cell methylomes isolated from human glioma tumor tissue. Using this data we generated a ranked list of the top hypo- and hypermethylated genomic regions and identified cell type specific clusters seen in different pathological states by looking at Differentially Methylated Regions (DMR) uncovering unique single-cell methylation profiles that may be obscured by bulk or pseudo-bulk analysis. These data show that the ScaleBio single-cell methylation workflow offers increased sensitivity, specificity, and accuracy in identifying DNA methylation sites when compared to other techniques while offering a comprehensive view of methylomes and providing insights into cellular heterogeneity and trajectories. Citation Format: Sanika Khare, Dominic Skinner, Maggie Nakamoto, Hosu Sin, Ashley Woodfin, Eric Pu, Lin Lin, Jason Koth, Beth Walczak. Enhanced single cell DNA methylation analysis using combinatorial indexing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(7_Suppl):Abstract nr LB193.

Optimal design and analysis of a grid‐connected hybrid renewable energy system using HOMER Pro: A case study of Tumbatu Island, Zanzibar

AbstractThis study addresses the pressing issue of quality electricity access in remote regions, with a specific focus on Tumbatu Island in Tanzania. Most studies on quality improvement concentrate on low‐voltage distribution lines and leave high‐voltage (HV) transmission lines behind. It is essential to address quality issues in HV lines due to their critical role in electricity transmission and distribution infrastructure. The objective of the study is to improve the voltage profile on the island's HV transmission line by identifying the optimal hybrid energy system comprising solar PV, wind turbine, and battery technologies. The study begins by presenting the total power demand and consumption on Tumbatu Island, which are important factors in designing an efficient energy system. The findings reveal a power demand of 7173 kW and a consumption of 28540 kWh/day, with an average scaled value of 1507.9 and 6000 kWh/day. To achieve the desired voltage profile improvement, the research incorporates HOMER Pro software for simulating and analyzing various hybrid system configurations. Essential input parameters, such as costs, resources, technology components, and load, are considered in the simulation process. The software generates a ranked list of options based on the system's net present cost (NPC). The simulation results demonstrate that the integration of solar PV, wind turbine, and HV lines provides a substantial improvement in the voltage profile, increasing it from 29.6 to 31.23 kV during peak demand periods. This solution proves to be the most economically viable, with the lowest NPC of $4,003,851 and a relatively short payback period of 3.79 years. In addition, a sensitivity analysis is performed to identify the most influential parameter in the system's performance. Wind speed is found to have the greatest impact, emphasizing its significance in the design and operation of the hybrid energy system. The implementation of this optimal hybrid renewable energy system on Tumbatu Island will not only improve the voltage profile and meet the island's energy needs but also shall contribute to global efforts in reducing pollution and the cost of electricity to the Tumbatu Island population. Moreover, it addresses the current and future demands for clean energy, underlining its importance in achieving sustainable and accessible electricity in sub‐Saharan Africa and beyond.

BootGSEA: a bootstrap and rank aggregation pipeline for multi-study and multi-omics enrichment analyses.

Introduction: Gene set enrichment analysis (GSEA) subsequent to differential expression analysis is a standard step in transcriptomics and proteomics data analysis. Although many tools for this step are available, the results are often difficult to reproduce because set annotations can change in the databases, that is, new features can be added or existing features can be removed. Finally, such changes in set compositions can have an impact on biological interpretation. Methods: We present bootGSEA, a novel computational pipeline, to study the robustness of GSEA. By repeating GSEA based on bootstrap samples, the variability and robustness of results can be studied. In our pipeline, not all genes or proteins are involved in the different bootstrap replicates of the analyses. Finally, we aggregate the ranks from the bootstrap replicates to obtain a score per gene set that shows whether it gains or loses evidence compared to the ranking of the standard GSEA. Rank aggregation is also used to combine GSEA results from different omics levels or from multiple independent studies at the same omics level. Results: By applying our approach to six independent cancer transcriptomics datasets, we showed that bootstrap GSEA can aid in the selection of more robust enriched gene sets. Additionally, we applied our approach to paired transcriptomics and proteomics data obtained from a mouse model of spinal muscular atrophy (SMA), a neurodegenerative and neurodevelopmental disease associated with multi-system involvement. After obtaining a robust ranking at both omics levels, both ranking lists were combined to aggregate the findings from the transcriptomics and proteomics results. Furthermore, we constructed the new R-package "bootGSEA," which implements the proposed methods and provides graphical views of the findings. Bootstrap-based GSEA was able in the example datasets to identify gene or protein sets that were less robust when the set composition changed during bootstrap analysis. Discussion: The rank aggregation step was useful for combining bootstrap results and making them comparable to the original findings on the single-omics level or for combining findings from multiple different omics levels.

Explainable artificial intelligence for feature selection in network traffic classification: A comparative study

AbstractOver the past decade, there has been a growing surge of interest in leveraging artificial intelligence and machine learning models to address real‐world challenges within the field of telecommunications and networking. Among these challenges, network traffic classification has consistently remained a key focal point within the community. The emergence of new paradigms such as internet of things has added complexity to this challenge, owing to the diverse array of traffic classes it introduces. This complexity is further compounded when dealing with encrypted traffic data. In the context of this paper, we employ explainable artificial intelligence (XAI) techniques to offer insightful explanations for the task of classifying encrypted network traffic. Our approach involves presenting a ranked list of the model's input features, organized based on their significance in influencing the classifier's predictions. These ranked lists shed light on the underlying functionality of the encrypted traffic classification model, enabling users to discern which features hold greater importance in shaping the model's output by harnessing the insights provided by XAI approaches. Moreover, we carry out a quantitative assessment and comparison of the results generated by these XAI methods and outputs of three other conventional, widely adopted feature selection approaches, utilizing them as feature selection techniques when applied to a convolutional neural network for classifying data from the CIC‐Darknet2020 and CIC‐IDS2017 datasets. The results show that XAI techniques have high performance in feature selection for network traffic classification.

Spectrum-based rule- and item-level localization of faults in context-free grammars

We describe and evaluate spectrum-based methods aimed at finding faults in context-free grammars. In their basic form, they take as input a test suite and a parser for the grammar that is modified to collect grammar spectra (i.e., the sets of grammar elements used in attempts to parse the individual test cases), and return as output a ranked list of suspicious elements. We define grammar spectra suitable for localizing faults on the level of the grammar rules (i.e., rule spectra) and the rules’ individual symbols (i.e., item spectra), respectively. We show how both types of grammar spectra can be collected by both LL and LR parsers, and how the JavaCC, ANTLR, and CUP parser generators can be modified and used to automate the collection of the grammar spectra. We also show how grammar spectra can be synthesized directly from test cases derived from a grammar, and how such synthetic spectra can be used to localize differences between a grammar and a black-box system under test.We first evaluate our approach over a large number of medium-sized single fault grammars, which we constructed by fault seeding from a common origin grammar. At the rule level, it ranks the rules containing the seeded faults within the top five rules in about 40%–70% of the cases, depending on the applied parsing technique, test suite, and ranking metric, and pinpoints them (i.e., correctly identifies them as unique most suspicious rule) in about 10%–30% of the cases, with significantly better results for the synthetic spectra. At the item level, our approach remains remarkably effective despite the larger number of possible locations, provided it is coupled with a simple tie-breaking strategy that prefers items with the right-most designated position over other items from the same rules in a tie. It typically ranks the seeded faults within the top five positions in about 30%–60% of the cases, and pinpoints them in about 15%–40% of the cases. This specialized item-level localization also significantly outperforms a simplistic extension of the rule-level localization, where all positions within a rule are given the same score.We further evaluate our approach over grammars that contain real faults. We show that an iterative method can be used to localize and manually remove one by one multiple faults in grammars submitted by students enrolled in various compiler engineering courses; in most iterations, the top-ranked rule already contains an error, and no error is ranked outside the top five ranked rules. We finally apply our approach to a large open-source SQLite grammar and show where this version deviates from the language accepted by the actual SQLite system.

Identification of top influence users in disseminating information on the 2024 Indonesian National Election

Social media has a vital role in general elections in Indonesia because social media is one of the platforms used by presidential candidates for campaigns to gain public support. General elections in Indonesia occur every five years. Many tweets talk about presidential candidates approaching the national election period. Not least, some buzzers deliberately use Twitter to carry out propaganda against a candidate or to bring down other presidential candidates with their opinions because information can spread widely and quickly on Twitter. Based on this, it is necessary to identify influential users in disseminating information related to the 2024 National Election, especially on Twitter. Various centrality methods were used in this study to identify influence users in sharing information about the 2024 National Election such us Degree Centrality, Closeness Centrality, Harmonic Centrality, Eigenvector Centrality, and Load Centrality. For the evaluation in this study, the results of each method were compared to one another to measure the similarity and correlation between the ranking lists of users who were influential in disseminating information about the 2024 National Election.

Measuring and Analyzing Scholarly Literature Published on Diabetes Mellitus Type 1 with Special Reference to Bradford Law of Scattering and Leimkuhler Model: A Scientometric Study

The present paper deals with the applicability of Bradford’s law of Scattering to the literature on Diabetes Mellitus type 1 disease. The PubMed database has been selected to collect data for the present study. A total of 489 journals published 1542 articles during the year 2012 to 2021 on Diabetes Mellitus type 1 disease. It was found that Current Diabetes Reports was the top core journal that published the highest number of articles (100) along with citation 4550. Bradford's law was tested for applicability using a ranked list of journals and publishing output. The Diabetes Mellitus type 1 literature's journal distribution pattern corresponded to Bradford's distribution pattern. The applicability of Leimkuhler’s model was also tested and it is also fitted to the selected disease dataset of Diabetes Mellitus Type-1

Vascular In-person for Students In the match Trial: An investigation of postinterview site visits to address the limitations of virtual interviews

PurposeThe graduate medical education community implemented virtual residency interviews in response to travel restrictions during the COVID-19 pandemic, and this approach has persisted. Although many residency applicants wish to visit in-person prospective training sites, such opportunities could bias programs toward those who are able to meet this financial burden, exacerbating equity concerns. One proposed solution is to offer applicants the opportunity to visit only after a program’s rank list is “locked,” avoiding favoritism to applicants who visit, but allowing applicants to experience some of the camaraderie, geography, and local effects of an in-person visit. As debate about the optimal format of residency interviews continues, it is important to investigate whether in-person program visits, completed after program rank list certification, provide meaningful benefits to applicants in the residency match process. MethodsAll vascular programs entering the 2023 integrated vascular surgery residency match were invited to participate. Programs agreed to certify their National Resident Matching Program rank lists by February 1, 2023. Applicants then had the opportunity to visit the programs at which they interviewed. The particulars of the visit were determined by the individual programs. Applicants completed their standard rank list and locked on the standard date: March 1, 2023. Applicants then completed a survey regarding the impact of the visits on their rank order list decision-making. Program directors (PDs) completed a survey regarding their experiences as well. Data were collected using REDCap. ResultsTwenty-one of the 74 (28%) programs participated. Nineteen PDs completed the postinterview site visit survey (response rate 90%). Applicants interviewing at the participating programs (n = 112) were informed of the study, offered the opportunity to attend postinterview site visits, and received the survey. Forty-seven applicants responded (response rate 42%). Eighty-six percent of applicants stated that the visit impacted their rank list. Most important factors were esprit de corps of the program (86%), the faculty/trainees/staff (81%), and the physical setting (62%). Seventy-one percent of those participating spent ≤$800 on their visit. Eighty-one percent were satisfied with the process. Twenty-one percent of PDs would have changed their rank list if they could have based on the applicants’ in-person visit. Sixty-three percent of the visit sessions cost the programs ≤$500, and 63% were satisfied with the process. ConclusionsThis study is the first to document the impact of in-person site visits by applicants on a graduate medical education match process in one specialty. Our results suggest that this process provides meaningful data to applicants that helped them with their decision-making evidenced by most altering their rank lists, while avoiding some of the critical equity issues that accompany traditional in-person interviews. This may provide a model for future interview processes for residency programs.

HBCR_DMR: A Hybrid Method Based on Beta-Binomial Bayesian Hierarchical Model and Combination of Ranking Method to Detect Differential Methylation Regions in Bisulfite Sequencing Data.

DNA methylation is a key epigenetic modification involved in gene regulation, contributing to both physiological and pathological conditions. For a more profound comprehension, it is essential to conduct a precise comparison of DNA methylation patterns between sample groups that represent distinct statuses. Analysis of differentially methylated regions (DMRs) using computational approaches can help uncover the precise relationships between these phenomena. This paper describes a hybrid model that combines the beta-binomial Bayesian hierarchical model with a combination of ranking methods known as HBCR_DMR. During the initial phase, we model the actual methylation proportions of the CpG sites (CpGs) within the replicates. This modeling is achieved through beta-binomial distribution, with parameters set by a group mean and a dispersion parameter. During the second stage, we establish the selection of distinguishing CpG sites based on their methylation status, employing multiple ranking techniques. Finally, we combine the ranking lists of differentially methylated CpG sites through a voting system. Our analyses, encompassing simulations and real data, reveal outstanding performance metrics, including a sensitivity of 0.72, specificity of 0.89, and an F1 score of 0.76, yielding an overall accuracy of 0.82 and an AUC of 0.94. These findings underscore HBCR_DMR's robust capacity to distinguish methylated regions, confirming its utility as a valuable tool for DNA methylation analysis.

Factors Influencing Underrepresented Minority Applicants’ Ranking of General Surgery Residency Programs

Background: Underrepresented minorities (URM) are disproportionately lacking in academic surgery, yet there is wide variation in URM in residency programs. This mixed methods study aims to describe factors that URM interns identified as important in their rank list of general surgery residency programs. These data will aid institutions in increasing URM representation and creating supportive training environments in academic surgery. Methods: We employed a mixed methods approach using an online survey and semi-structured interviews developed through focus group engagement and literature review. Interviews were audio recorded, transcribed, and coded for themes. We conducted narrative thematic analysis. Results: Sixty-four incoming general surgery interns completed the survey with median age 25 to 29. Of respondents, 67% identified as female and 34% as URM. URM general surgery interns were more likely to rate program benefits and finances, revisit opportunities (ie, organized post-interview revisit weekends), and post-graduate job opportunities higher than were non-URM participants (p<0.05). URM participants also ranked factors such as diversity of residents and faculty and diversity and inclusion statements by programs higher than did non-URM participants (p<0.05). Five URM interns completed semi-structured interviews. Qualitative themes influencing URM interns’ rank list decisions included (1) presence of URM in the program, (2) program support of resident-led goals, (3) active inclusivity initiatives, and (4) direct experiences of current URM residents. All interview participants relied heavily on interactions with current URM residents as a source of information. Detractors from programs included interns concerned with being the only person of color at their prospective program and no concrete initiatives or goals in place to support diversity and inclusion, raising concern for performative allyship. Conclusions: When evaluating residency programs, URM weighed current representation and support of URMs, active diversity and inclusion initiatives within a program, and lived experiences of URM residents. Surgery residency programs should take steps to not only intentionally recruit URM applicants but also create a supportive training environment that is diverse, inclusive, and empowering.

Distribution of grades and rank lists among Emergency Medicine programs during the 2022-2023 academic year

Distribution of grades and rank lists among Emergency Medicine programs during the 2022-2023 academic year

Learning to Rank in Generative Retrieval

Generative retrieval stands out as a promising new paradigm in text retrieval that aims to generate identifier strings of relevant passages as the retrieval target. This generative paradigm taps into powerful generative language models, distinct from traditional sparse or dense retrieval methods. However, only learning to generate is insufficient for generative retrieval. Generative retrieval learns to generate identifiers of relevant passages as an intermediate goal and then converts predicted identifiers into the final passage rank list. The disconnect between the learning objective of autoregressive models and the desired passage ranking target leads to a learning gap. To bridge this gap, we propose a learning-to-rank framework for generative retrieval, dubbed LTRGR. LTRGR enables generative retrieval to learn to rank passages directly, optimizing the autoregressive model toward the final passage ranking target via a rank loss. This framework only requires an additional learning-to-rank training phase to enhance current generative retrieval systems and does not add any burden to the inference stage. We conducted experiments on three public benchmarks, and the results demonstrate that LTRGR achieves state-of-the-art performance among generative retrieval methods. The code and checkpoints are released at https://github.com/liyongqi67/LTRGR.

Abstract 2330: CTD2 "Connects the Dots" to capture disease genes in complex networks and its application on the Cancer Genomics Cloud

Abstract We have developed CTD2 (an algorithm to "Connect the Dots") to capture biological signals and identify candidate genes in complex networks. CTD2 extends the functionality of its predecessor CTD by providing ranked lists of genes that are “guilty by association” and significantly linked to other genes of interest. With the explosion of large scale multi-omics studies in the cancer field, novel approaches are needed to interpret these valuable datasets. Both CTD and CTD2 are information-theoretic algorithms that allow identification of highly connected sets of genes in complex networks without the need for permutation testing. CTD has been previously used to interpret perturbations in different subtypes of breast cancer. Additionally CTD has been used to identify biomarkers of chemotherapy response in Triple Negative Breast Cancer (TNBC) murine PDX models to both platinum and taxane agents. These small multigene biomarkers of response were shown to be informative for the response of both patients and PDXs. CTD2 was developed to expand the utility of the CTD package by also capturing genes that are "guilty by association". These genes are significantly connected to genes of interest (such as disease genes), and are ranked by their connectedness to these informative gene sets. To demonstrate its utility, we investigated if CTD2 could identify known breast cancer genes. Using TCGA breast cancer expression data, we built case/control graphs over 5,000 variable genes for each subtype of breast cancer. Genes previously associated with breast cancer were identified with DisGeNET and split into a training set discovered pre-2015 and test set discovered post-2015. The genes that were discovered pre-2015 that overlapped with our networks (n = 680) were then used as an input for CTD2 along with the case/control graphs. We then ranked the connectedness of all the genes in these graphs to the training set and found that the test set of breast cancer genes that were discovered post-2015 were significantly enriched in these ranked lists.We have shown that CTD2 and CTD can be utilized to discover biologically informative signals in complex networks. Furthermore we have deployed these tools on the Cancer Genomics Cloud to make them easily accessible for users without a bioinformatics background. The democratization of these in silico tools will allow for their adaptation by a wider audience and aid in the interpretation of large multi-omic datasets. Citation Format: Varduhi Petrosyan, Vladimir Kovacevic, Predrag Obradovic, Cera Fisher, Zelia Worman, Divya Sain, Jack DiGiovanna, Brandi Davis-Dusenberry, Aleksandar Milosavljevic. CTD2 "Connects the Dots" to capture disease genes in complex networks and its application on the Cancer Genomics Cloud [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2330.

Abstract 4493: Protein structure inspired drug discovery

Abstract Introduction: Drug action is mediated by a small molecule therapeutic binding to a specific structural motif on a protein. As structure determines function, it follows that unique structural motifs are enriched for sites of unique function and high therapeutic target value. We hypothesized that protein structure constitutes high value primary information that can drive the discovery of novel therapeutic agents. Method: We developed a unique physics-based protein structure analysis platform scaled to run on a supercomputer, allowing us to consider dynamic protein flexibility and the influence of solvent upon protein structure, and to probe large protein libraries as a primary source of structure information that can support up-front structure-based screens. We developed and then probed a high resolution protein x-ray crystallographic library, created and curated by us. From a 60 million compound library, we conducted virtual screens for docking to identified sites. 8-day colony formation assays were conducted on a panel of 8 human breast, prostate, lung and colon cancer cell lines, as well as 8- and 14-day trilineage hematopoietic colony formation assays on human cord blood stem cells. Findings: From the first 180 deposited protein structures, we screened for unique protein surface structure. Structures were deemed unique if they were not otherwise known to be associated with any particular functional roles and were not the binding site of known therapeutics. Further selection included those that were clefts and possessed physical characteristics compatible with binding small chemicals whose properties have been linked to effective therapeutics. For each of 8 sites on 6 different proteins, a suite of analytics probed docking solutions from the compound library, generating a ranked list of 100 compounds for each site. Based on factors inclusive of availability and low potential for toxicity, compounds were acquired and used in in-parallel selection and de-selection functional assays. Cancer cell growth inhibition was a positive selection criterion, with additional ranking based on greater potency and broader efficacy. Bone marrow toxicity is limiting for most drugs, and inhibition of bone marrow stem cell growth was as a deselection criterion. Multiple compounds clustered around 2 sites on 2 different proteins. One compound, Dxr2-017, exhibited very high relative efficacy against colon and breast cancer cells. Evaluation of Dxr2-017 in the NCI-60 panel corroborated these findings, and showed even higher efficacy against melanoma. We demonstrated that Dxr2-017 inhibited melanoma growth with low nanomolar efficacy, while concentrations of over 2300-fold higher had only limited inhibitory effects on bone marrow stem cells. Conclusions: This study provides proof-of-principle that protein structure constitutes high value primary information that can support identification of novel therapeutics. This approach is widely applicable and expandable. Citation Format: Fangfang Qiao, T. Andrew Binkowski, Wayne Anderson, Weining Chen, Gray Schiltz, Karl Scheidt, Amarnath Natarajan, Raymond Bergan. Protein structure inspired drug discovery [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4493.

Abstract 4685: The identification of new therapeutic targets in mucinous ovarian carcinoma

Abstract Background: Epithelial ovarian cancer (EOC) is the second cause of death among gynaecological cancersworldwide. Mucinous ovarian carcinoma (mEOC) accounts for 3-5% of all EOCs, and when diagnosed at anadvanced stage its prognosis is very poor mainly for its limited chemosensitivity. PLK1 is a member of thewell-conserved serine/threonine kinase family, that plays a key role in the progression of mitosis, in the G2/Mcheckpoint regulation, DNA damage replication, stress response, and cell death pathways. mEOC has beenshown to be sensitive to PLK1 inhibition both with siRNA and small molecule inhibitor (onvansertib). Giventhis background and the need for new therapeutic approaches in mEOC, the purposes of this work were toidentify new therapeutic targets in mEOC using CRISPR/Cas9 lentivirus libraries and to find synthetic lethalpartners with the PLK1 inhibitor onvansertib. Methods: We used three different mucinous cell lines: MCAS,EFO27, TOV2414, overexpressing Cas9 gene. We evaluated the Cas9 expression through Western Blotanalysis, and the Cas9 activity through a Fluorescent Activated Cell Sorter- based-GFP assay. We used twodifferent lentiviral gRNAs Libraries (Bassik Library Human CRISPR Deletion Library - Apoptosis and cancer(Addgene #101926)/Drug targets, kinases, and phosphatases (Addgene #101927)). The EFO27 Cas9 cells wereused for the screening experiment. Cells were infected with the lentiviral libraries at MOI of 0.3 and put inpuromycin selection. Six days after puromycin addition (Time point 0) 30x10^6 cells were collected forsequencing. The remaining 30x10^6 cells were split in two: control cells and cells treated with a subtoxic doseof onvansertib (IC30). After 7 days (Time point 1) 30x10^6 cells for each condition were collected forsequencing. The infection of libraries was performed alone and in combination with sub-toxic onvansertibtreatment, to identify genes important for mucinous carcinoma cell survival, and those genes in syntheticlethality with onvansertib. Results: Bioinformatic analyses comparing control cells (time 1 vs time 0) andPLK1 treated and non -treated cells (time 1) allowed the generation of a gene ranking list, including genesinvolved in survival of EFO27 cells like ZC2HC1C, RPA2, POLE3, KIN, TUBG1, SMC2, CDC26, CDC42, HOXA9,TAF10, DCLRE1C, SENP1, MRPS31, COPS2; and genes in synthetic lethality with PLK1 treatment, like JUND,CARD9, BCL2L2. Validation experiments are ongoing and will be presented and discussed. Conclusion: Using a CRISPR-based approach, we identified genes that either alone or in combination with PLK inhibitioncould have therapeutic value in mEOC. Citation Format: Serena Petrella, Marika Colombo, Mirko Marabese, Chiara Grasselli, Andrea Panfili, Ilaria Craparotta, Maria Chiara Barbera, Giada Andrea Cassanmagnago, Marco Bolis, Giovanna Damia. The identification of new therapeutic targets in mucinous ovarian carcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4685.

Enhancing Natural Language Processing (NLP) through VIKOR Method: A Comprehensive Approach for Improved Computational Linguistics

Natural Language Processing (NLP) is a revolutionary discipline that resides at the crossroads of computer science and linguistics, with its primary emphasis on the interaction between computers and human language. With roots in artificial intelligence, NLP seeks to equip machines with the ability to comprehend, interpret, and respond to natural language, enabling more intuitive and meaningful communication between humans and computers. This multidisciplinary domain leverages advanced algorithms and models to tackle a range of linguistic challenges, from language translation and sentiment analysis to speech recognition. Recent breakthroughs, particularly in deep learning and neural networks, have propelled NLP to new heights, with applications spanning diverse sectors such as healthcare, finance, and education. As NLP continues to advance, its potential impact on enhancing human-machine interaction and information processing is increasingly evident, promising a future where technology seamlessly integrates with our natural modes of communication. The significance of Natural Language Processing (NLP) in research lies in its capacity to revolutionize human-computer interaction. NLP empowers machines to understand and generate human language, enabling advancements in sentiment analysis, language translation, and conversational AI. This transformative capability has profound implications across various industries, including healthcare, finance, and education. NLP research not only enhances the efficiency of information retrieval and processing but also holds the key to developing more intuitive and user-friendly technologies. As the frontier of NLP research expands, its potential to bridge the communication gap between humans and machines continues to shape the future of technology and information access. VIKOR functions as a multi-criteria decision-making technique that determines the best option by comparing it to the ideal option. The ranking procedure comprises determining distances concerning the optimal solution. VIKOR uses linear normalization to get the best possible outcomes. This technique was first presented by Opricovic in 1998 and was intended to optimize multi-attribute complex systems. Its main focus is on ranking lists that allow for flexibility in the strategy weight interval and incorporate compromise alternatives to obtain desired results. Alternative taken as Open AI GPT-4, Google BERT, Microsoft Azure Text Analytics, IBM Watson Natural Language Understanding, spaCy, NLTK (Natural Language Toolkit), Amazon Comprehend. Evaluation Parameter taken as Sentiment Analysis (F1 Score), NER Performance, Language Support, Processing Speed (Response Time) (ms), Translation Accuracy (BLEU Score), Data Privacy. The result it is seen that Amazon Comprehend is got the first rank where as is the IBM Watson Natural Language Understanding is having the lowest rank.

The Expertise Finder: Using NASA ADS for Panel Reviewer Selection

Astronomical research is often made possible by the utilization of different resources that support members of the community in their research endeavors (e.g., telescopes access, or funds). Often, the entities that provide these resources establish a competitive process, which include the submission, by interested parties, of research proposals describing the research objectives and how the resources will be used to reach them. These proposals are evaluated by “ad hoc” panels of researchers, that provide a ranked list of the proposals based on their scientific relevance and other factors. This is also true for many NASA funding opportunities. In this article we present an open-sourced tool, the Expertise Finder, that simplifies the reviewers identification process. Expertise Finder provides NASA officials ways to scout prospective reviewers at small, less resourced, institutions that may have fewer opportunities to present and publish their research work with respect to colleagues at larger institutions.

Towards Simulation-Based Evaluation of Recommender Systems with Carousel Interfaces

Offline data-driven evaluation is considered a low-cost and more accessible alternative to the online empirical method of assessing the quality of recommender systems. Despite their popularity and effectiveness, most data-driven approaches are unsuitable for evaluating interactive recommender systems. In this article, we attempt to address this issue by simulating the user interactions with the system as a part of the evaluation process. Particularly, we demonstrate that simulated users find their desired item more efficiently when recommendations are presented as a list of carousels compared to a simple ranked list.