The causes of the reproducibility crisis include lack of standardization and transparency in scientific reporting. Checklists such as ARRIVE and CONSORT seek to improve transparency, but they are not always followed by authors and peer review often fails to identify missing items. To address these issues, there are several automated tools that have been designed to check different rigor criteria. We have conducted a broad comparison of 11 automated tools across 9 different rigor criteria from the ScreenIT group. We found some criteria, including detecting open data, where the combination of tools showed a clear winner, a tool which performed much better than other tools. In other cases, including detection of inclusion and exclusion criteria, the combination of tools exceeded the performance of any one tool. We also identified key areas where tool developers should focus their effort to make their tool maximally useful. We conclude with a set of insights and recommendations for stakeholders in the development of rigor and transparency detection tools. The code and data for the study is available at https://github.com/PeterEckmann1/tool-comparison.

Advances in large language models (LLMs) have recently opened new and promising avenues for small-molecule drug discovery. Yet existing LLM-based approaches for molecular generation often suffer from high rates of invalid and low-quality ligand candidates, a result of the syntactic limitations of current models with regard to molecular strings. In this paper, we introduce , an evolutionary agentic framework for de novo drug design. combines a multi-objective genetic algorithm with an agentic LLM operator that iteratively updates the ligand population. We build a comprehensive toolbox of RDKit-backed functions that allows our agentic operator to consisently make precise ligand modifications. achieves state-of-the-art performance on multi-objective property optimization tasks, discovering drug-like and synthesizable ligands that have >10% stronger predicted binding affinity compared to existing methods, evaluated on three protein targets. ligands additionally achieve state-of-the-art results in gold-standard Absolute Binding Free Energy scores, gaining over existing methods by over 35%. By studying chain-of-thought reasoning traces, we observe that tool-calling enables the model to more faithfully execute its planned modifications, efficiently exploiting the strong chemical prior knowledge in LLMs.

Current generative models for drug discovery primarily use molecular docking as an oracle to guide the generation of active compounds. However, such models are often not useful in practice because even compounds with high docking scores do not consistently show real-world experimental activity. More accurate methods for activity prediction exist, such as molecular dynamics based binding free energy calculations, but they are too computationally expensive to use in a generative model. To address this challenge, we propose Multi-Fidelity Latent space Active Learning (MF-LAL), a generative modeling framework that integrates a set of oracles with varying cost-accuracy tradeoffs. Using active learning, we train a surrogate model for each oracle and use these surrogates to guide generation of compounds with high predicted activity. Unlike previous approaches that separately learn the surrogate model and generative model, MF-LAL combines the generative and multi-fidelity surrogate models into a single framework, allowing for more accurate activity prediction and higher quality samples. Our experiments on two disease-relevant proteins show that MF-LAL produces compounds with significantly better binding free energy scores than other single and multi-fidelity approaches (~50% improvement in mean binding free energy score). The code is available at https://github.com/Rose-STL-Lab/MF-LAL.

Predicting the activities of new compounds against biophysical or phenotypic assays based on the known activities of one or a few existing compounds is a common goal in early stage drug discovery. This problem can be cast as a "few-shot learning" challenge, and prior studies have developed few-shot learning methods to classify compounds as active versus inactive. However, the ability to go beyond classification and rank compounds by expected affinity is more valuable. We describe Few-Shot Compound Activity Prediction (FS-CAP), a novel neural architecture trained on a large bioactivity data set to predict compound activities against an assay outside the training set, based on only the activities of a few known compounds against the same assay. Our model aggregates encodings generated from the known compounds and their activities to capture assay information and uses a separate encoder for the new compound whose activity is to be predicted. The new method provides encouraging results relative to traditional chemical-similarity-based techniques as well as other state-of-the-art few-shot learning methods in tests on a variety of ligand-based drug discovery settings and data sets. The code for FS-CAP is available at https://github.com/Rose-STL-Lab/FS-CAP.

Current generative models for drug discovery primarily use molecular docking to evaluate the quality of generated compounds. However, such models are often not useful in practice because even compounds with high docking scores do not consistently show experimental activity. More accurate methods for activity prediction exist, such as molecular dynamics based binding free energy calculations, but they are too computationally expensive to use in a generative model. We propose a multi-fidelity approach, Multi-Fidelity Bind (MFBind), to achieve the optimal trade-off between accuracy and computational cost. MFBind integrates docking and binding free energy simulators to train a multi-fidelity deep surrogate model with active learning. Our deep surrogate model utilizes a pretraining technique and linear prediction heads to efficiently fit small amounts of high-fidelity data. We perform extensive experiments and show that MFBind (1) outperforms other state-of-the-art single and multi-fidelity baselines in surrogate modeling, and (2) boosts the performance of generative models with markedly higher quality compounds.

This technical report investigates variants of the Latent Inceptionism on Molecules (LIMO) framework to improve the properties of generated molecules. We conduct ablative studies of molecular representation, decoder model, and surrogate model training scheme. The experiments suggest that an autogressive Transformer decoder with GroupSELFIES achieves the best average properties for the random generation task.

Predicting the activities of compounds against protein-based or phenotypic assays using only a few known compounds and their activities is a common task in target-free drug discovery. Existing few-shot learning approaches are limited to predicting binary labels (active/inactive). However, in real-world drug discovery, degrees of compound activity are highly relevant. We study Few-Shot Compound Activity Prediction (FS-CAP) and design a novel neural architecture to meta-learn continuous compound activities across large bioactivity datasets. Our model aggregates encodings generated from the known compounds and their activities to capture assay information. We also introduce a separate encoder for the unknown compound. We show that FS-CAP surpasses traditional similarity-based techniques as well as other state of the art few-shot learning methods on a variety of target-free drug discovery settings and datasets.

Abstract Antibodies are ubiquitous key biological research resources yet are tricky to use as they are prone to performance issues and represent a major source of variability across studies. Understanding what antibody was used in a published study is therefore necessary to repeat and/or interpret a given study. However, antibody reagents are still frequently not cited with sufficient detail to determine which antibody was used in experiments. The Antibody Registry is a public, open database that enables citation of antibodies by providing a persistent record for any antibody-based reagent used in a publication. The registry is the authority for antibody Research Resource Identifiers, or RRIDs, which are requested or required by hundreds of journals seeking to improve the citation of these key resources. The registry is the most comprehensive listing of persistently identified antibody reagents used in the scientific literature. Data contributors span individual authors who use antibodies to antibody companies, which provide their entire catalogs including discontinued items. Unlike many commercial antibody listing sites which tend to remove reagents no longer sold, registry records persist, providing an interface between a fast-moving commercial marketplace and the static scientific literature. The Antibody Registry (RRID:SCR_006397) https://antibodyregistry.org.

Scoring ligands in a library based on their structural similarity to a known hit compound is widely used in drug discovery following high-throughput screening. However, such "similarity search" relies on the assumption that structurally similar compounds have similar activities, and will therefore only retrieve ligands with hit-like affinity, requiring resource-intensive tweaking by medicinal chemists to reach a more active lead compound. We propose a novel approach, One-Shot Ligand Scoring (OSLS), that is much more capable of directly retrieving lead-like compounds from a library using a novel one-shot learning technique. For this new task, we design a Siamese-inspired neural architecture using two Transformer encoders without tied weights, a novel positional encoding-like mechanism, and a final prediction head. OSLS is able to score ligands by activity against a target without any target-specific knowledge beyond a single known activity value, a cost-effective approach to ligand-based or phenotypic drug discovery. We show that OSLS surpasses traditional similarity search as well as modern deep learning baselines on a simulated ligand retrieval task. Furthermore, we demonstrate the applicability of our approach on various drug discovery tasks that also involve ligand scoring, including drug repositioning, precision patient-level drug efficacy prediction, and even molecular generative modeling.

Antibodies are critical reagents to detect and characterize proteins. It is commonly understood that many commercial antibodies do not recognize their intended targets, but information on the scope of the problem remains largely anecdotal, and as such, feasibility of the goal of at least one potent and specific antibody targeting each protein in a proteome cannot be assessed. Focusing on antibodies for human proteins, we have scaled a standardized characterization approach using parental and knockout cell lines (Laflamme et al., 2019) to assess the performance of 614 commercial antibodies for 65 neuroscience-related proteins. Side-by-side comparisons of all antibodies against each target, obtained from multiple commercial partners, have demonstrated that: (i) more than 50% of all antibodies failed in one or more applications, (ii) yet, ~50–75% of the protein set was covered by at least one high-performing antibody, depending on application, suggesting that coverage of human proteins by commercial antibodies is significant; and (iii) recombinant antibodies performed better than monoclonal or polyclonal antibodies. The hundreds of underperforming antibodies identified in this study were found to have been used in a large number of published articles, which should raise alarm. Encouragingly, more than half of the underperforming commercial antibodies were reassessed by the manufacturers, and many had alterations to their recommended usage or were removed from the market. This first study helps demonstrate the scale of the antibody specificity problem but also suggests an efficient strategy toward achieving coverage of the human proteome; mine the existing commercial antibody repertoire, and use the data to focus new renewable antibody generation efforts.

Preprints, versions of scientific manuscripts that precede peer review, are growing in popularity. They offer an opportunity to democratize and accelerate research, as they have no publication costs or a lengthy peer review process. Preprints are often later published in peer-reviewed venues, but these publications and the original preprints are frequently not linked in any way. To this end, we developed a tool, PreprintMatch, to find matches between preprints and their corresponding published papers, if they exist. This tool outperforms existing techniques to match preprints and papers, both on matching performance and speed. PreprintMatch was applied to search for matches between preprints (from bioRxiv and medRxiv), and PubMed. The preliminary nature of preprints offers a unique perspective into scientific projects at a relatively early stage, and with better matching between preprint and paper, we explored questions related to research inequity. We found that preprints from low income countries are published as peer-reviewed papers at a lower rate than high income countries (39.6% and 61.1%, respectively), and our data is consistent with previous work that cite a lack of resources, lack of stability, and policy choices to explain this discrepancy. Preprints from low income countries were also found to be published quicker (178 vs 203 days) and with less title, abstract, and author similarity to the published version compared to high income countries. Low income countries add more authors from the preprint to the published version than high income countries (0.42 authors vs 0.32, respectively), a practice that is significantly more frequent in China compared to similar countries. Finally, we find that some publishers publish work with authors from lower income countries more frequently than others.

Background: Improving rigor and transparency measures should lead to improvements in reproducibility across the scientific literature; however, the assessment of measures of transparency tends to be very difficult if performed manually. Objective: This study addresses the enhancement of the Rigor and Transparency Index (RTI, version 2.0), which attempts to automatically assess the rigor and transparency of journals, institutions, and countries using manuscripts scored on criteria found in reproducibility guidelines (eg, Materials Design, Analysis, and Reporting checklist criteria). Methods: The RTI tracks 27 entity types using natural language processing techniques such as Bidirectional Long Short-term Memory Conditional Random Field-based models and regular expressions; this allowed us to assess over 2 million papers accessed through PubMed Central. Results: Between 1997 and 2020 (where data were readily available in our data set), rigor and transparency measures showed general improvement (RTI 2.29 to 4.13), suggesting that authors are taking the need for improved reporting seriously. The top-scoring journals in 2020 were the Journal of Neurochemistry (6.23), British Journal of Pharmacology (6.07), and Nature Neuroscience (5.93). We extracted the institution and country of origin from the author affiliations to expand our analysis beyond journals. Among institutions publishing >1000 papers in 2020 (in the PubMed Central open access set), Capital Medical University (4.75), Yonsei University (4.58), and University of Copenhagen (4.53) were the top performers in terms of RTI. In country-level performance, we found that Ethiopia and Norway consistently topped the RTI charts of countries with 100 or more papers per year. In addition, we tested our assumption that the RTI may serve as a reliable proxy for scientific replicability (ie, a high RTI represents papers containing sufficient information for replication efforts). Using work by the Reproducibility Project: Cancer Biology, we determined that replication papers (RTI 7.61, SD 0.78) scored significantly higher (P<.001) than the original papers (RTI 3.39, SD 1.12), which according to the project required additional information from authors to begin replication efforts. Conclusions: These results align with our view that RTI may serve as a reliable proxy for scientific replicability. Unfortunately, RTI measures for journals, institutions, and countries fall short of the replicated paper average. If we consider the RTI of these replication studies as a target for future manuscripts, more work will be needed to ensure that the average manuscript contains sufficient information for replication attempts.

The rising rate of preprints and publications, combined with persistent inadequate reporting practices and problems with study design and execution, have strained the traditional peer review system. Automated screening tools could potentially enhance peer review by helping authors, journal editors, and reviewers to identify beneficial practices and common problems in preprints or submitted manuscripts. Tools can screen many papers quickly, and may be particularly helpful in assessing compliance with journal policies and with straightforward items in reporting guidelines. However, existing tools cannot understand or interpret the paper in the context of the scientific literature. Tools cannot yet determine whether the methods used are suitable to answer the research question, or whether the data support the authors’ conclusions. Editors and peer reviewers are essential for assessing journal fit and the overall quality of a paper, including the experimental design, the soundness of the study’s conclusions, potential impact and innovation. Automated screening tools cannot replace peer review, but may aid authors, reviewers, and editors in improving scientific papers. Strategies for responsible use of automated tools in peer review may include setting performance criteria for tools, transparently reporting tool performance and use, and training users to interpret reports.

Abstract   This year marks the fiftieth anniversary of Archie Cochrane's seminal report “Effectiveness and Efficiency: random reflections on health services”, which spurred a paradigm shift in basing health policy and practice on robust empirical evidence. Indeed, evidence-informed decision- and policy-making has since permeated different levels of healthcare, public health, and health systems. More recently, the notion of “evidence-based” medicine has seen a shift towards “evidence-informed” health policy and practice, explicitly acknowledging that research evidence is but one of several factors influencing decision-making. In patient care, clinical expertise and a patient's own values and beliefs influence treatment and care; while in health policy-making, political and social factors, financial concerns, timing, and stakeholder relationships are decisive to successful implementation. Nonetheless, evidence-informed decision-making in health revolves around several core tenets, including a clear and well-communicated definition of objectives, using best-available, empirical evidence sourced from multiple sources, excellence and interdisciplinarity in the methods of analysis, and active collaboration across sectors and stakeholders to drive change. Across healthcare, public health, health systems and global health, actors continue to adapt these core principles into increasingly sophisticated strategies of evidence-informed decision-making. In the shadow of the global COVID-19 pandemic, exchanging best practices across these areas of application within the complex evidence ecosystem is a key driver for building more sustainable capacities in evidence-informed decision-making and partnerships to accelerate progress towards strengthening health systems and achieving global health equity. This panel aims to bring together major actors in knowledge translation and evidence-to-policy processes, highlighting latest resources and their relevance to public health, and stimulating collaboration and future action. With brief inputs from knowledge brokers, such as the European Observatory on Health Systems and Policies and the World Health Organization's (WHO) Evidence-informed Policy Network (EVIPNet), the panel will reflect on the current state of evidence-informed policy- and decision-making, showcase latest tools and resources, and discuss them from the specific perspectives of public health, coverage (with a focus on Health Technology Assessment), and patient and citizen involvement (with a focus on shared decision-making). Contributions from the audience will be sought in an open discussion round following initial inputs from panelists, with the objective of further shaping collaboration with involved stakeholders towards improving population health. Key messages • Building on current best practices in evidence-informed decision-making is crucial to achieving better outcomes in healthcare, public health and health systems. • Leveraging synergies and collaborating between different areas of application of evidence-informed decision-making can help ensure maximum impact and strengthen the evidence ecosystem. Speakers/Panellists Suszy Lessof European Observatory On Health Systems and Policies, Brussels, Belgium Tanja Kuchenmueller WHO, Geneva, Switzerland Elena Petelos CSFM & HSR-PH Lab, Faculty of Medicine, University of Crete, Iraklion, Greece Ansgar Gerhardus German Public Health Association - DGPH, Bremen, Germany Isabelle Scholl University Clinic Hamburg-Eppendorf, Hamburg, Germany

Abstract Preprints occupied the spotlight early in the pandemic, as scientists, the media and the public sought information on the evolving pandemic. While some in the scientific community embraced this shift, others were concerned about the quality of these papers, which had not yet undergone peer review. Furthermore, the flood of COVID-19 preprints quickly overwhelmed the scientific community's ability to monitor and assess new preprints. Automated screening tools that detect beneficial practices, or common problems, in preprints are one potential solution to this problem. These tools could potentially provide individualized feedback, allowing authors to improve their manuscripts prior to publication in a peer-reviewed journal. We have combined many tools into a single pipeline, called ScreenIT. ScreenIT assess factors such as open data and open code, blinding, randomization, power calculations, limitations sections, and data visualization problems. Since June 2020, we have used ScreenIT to screen and post daily reports on more than 23,000 new COVID-19 preprints deposited on bioRxiv and medRxiv. Results show that practices such as sharing data and code are relatively uncommon. Sample size calculations, blinding and randomization are rarely reported and most papers do not report the sex of participants, animals or samples. This work demonstrates the feasibility of using automated tools to rapidly screen many preprints in real time, and provide authors and readers with rapid feedback. However, this approach has important limitations. Automated screening tools can make mistakes. Tools can't always determine whether an item is relevant to a particular manuscript. Further studies are needed to determine whether feedback from automated tools is effective in encouraging authors to improve reporting.

Generation of drug-like molecules with high binding affinity to target proteins remains a difficult and resource-intensive task in drug discovery. Existing approaches primarily employ reinforcement learning, Markov sampling, or deep generative models guided by Gaussian processes, which can be prohibitively slow when generating molecules with high binding affinity calculated by computationally-expensive physics-based methods. We present Latent Inceptionism on Molecules (LIMO), which significantly accelerates molecule generation with an inceptionism-like technique. LIMO employs a variational autoencoder-generated latent space and property prediction by two neural networks in sequence to enable faster gradient-based reverse-optimization of molecular properties. Comprehensive experiments show that LIMO performs competitively on benchmark tasks and markedly outperforms state-of-the-art techniques on the novel task of generating drug-like compounds with high binding affinity, reaching nanomolar range against two protein targets. We corroborate these docking-based results with more accurate molecular dynamics-based calculations of absolute binding free energy and show that one of our generated drug-like compounds has a predicted KD (a measure of binding affinity) of 6 · 10-14 M against the human estrogen receptor, well beyond the affinities of typical early-stage drug candidates and most FDA-approved drugs to their respective targets. Code is available at https://github.com/Rose-STL-Lab/LIMO.

Preprints, versions of scientific manuscripts that precede peer review, are growing in popularity. They offer an opportunity to democratize and accelerate research, as they have no publication costs or a lengthy peer review process. Preprints are often later published in peer-reviewed venues, but these publications and the original preprints are frequently not linked in any way. To this end, we developed a tool, PreprintMatch, to find matches between preprints and their corresponding published papers, if they exist. This tool outperforms existing techniques to match preprints and papers, both on matching performance and speed. PreprintMatch was applied to search for matches between preprints (from bioRxiv and medRxiv), and PubMed. The preliminary nature of preprints offers a unique perspective into scientific projects at a relatively early stage, and with better matching between preprint and paper, we explored questions related to research inequity. We found that preprints from low income countries are published as peer-reviewed papers at a lower rate than high income countries (39.6% and 61.1%, respectively), and our data is consistent with previous work that cite a lack of resources, lack of stability, and policy choices to explain this discrepancy. Preprints from low income countries were also found to be published quicker (178 vs 203 days) and with less title, abstract, and author similarity to the published version compared to high income countries. Low income countries add more authors from the preprint to the published version than high income countries (0.42 authors vs 0.32, respectively), a practice that is significantly more frequent in China compared to similar countries. Finally, we find that some publishers publish work with authors from lower income countries more frequently than others. PreprintMatch is available at .