In the past decade, a surge in the amount of electronic health record (EHR) data in the United States has been attributed to a favorable policy environment created by the Health Information Technology for Economic and Clinical Health (HITECH) Act of 2009 and the 21st Century Cures Act of 2016. Clinical notes for patients’ assessments, diagnoses, and treatments are captured in these EHRs in free-form text by physicians, who spend a considerable amount of time entering and editing them. Manually writing clinical notes takes a considerable amount of a doctor’s valuable time, increasing the patient’s waiting time and possibly delaying diagnoses. Large language models (LLMs) possess the ability to generate news articles that closely resemble human-written ones. We investigate the usage of Chain-of-Thought (CoT) prompt engineering to improve the LLM’s response in clinical note generation. In our prompts, we use as input International Classification of Diseases codes and basic patient information. We investigate a strategy that combines the traditional CoT with semantic search results to improve the quality of generated clinical notes. Additionally, we infuse a knowledge graph built from clinical ontology to further enrich the domain-specific knowledge of generated clinical notes. We test our prompting technique on six clinical cases from the CodiEsp test dataset using GPT-4, and our results show that it outperformed the clinical notes generated by standard one-shot prompts.

Phylogenetic inference and reconstruction methods generate hypotheses on evolutionary history. Competing inference methods are frequently used, and the evaluation of the generated hypotheses is achieved using tree comparison costs. The Robinson–Foulds (RF) distance is a widely used cost to compare the topology of two trees, but this cost is sensitive to tree error and can overestimate tree differences. To overcome this limitation, a refined version of the RF distance called the Cluster Affinity (CA) distance was introduced. However, CA distances are symmetric and cannot compare different types of trees. These asymmetric comparisons occur when gene trees are compared with species trees, when disparate datasets are integrated into a supertree, or when tree comparison measures are used to infer a phylogenetic network. In this study, we introduce a relaxation of the original Affinity distance to compare heterogeneous trees called the asymmetric CA cost. We also develop a biologically interpretable cost, the Cluster Support cost that normalizes by cluster size across gene trees. The characteristics of these costs are similar to the symmetric CA cost. We describe efficient algorithms, derive the exact diameters, and use these to standardize the cost to be applicable in practice. These costs provide objective, fine-scale, and biologically interpretable values that can assess differences and similarities between phylogenetic trees.

Although many user-friendly workflows exist for identifications of peptides and proteins in mass-spectrometry-based proteomics, there is a need of easy to use, fast, and accurate workflows for identifications of microorganisms, antimicrobial resistant proteins, and biomass estimation. Identification of microorganisms is a computationally demanding task that requires querying thousands of MS/MS spectra in a database containing thousands to tens of thousands of microorganisms. Existing software can't handle such a task in a time efficient manner, taking hours to process a single MS/MS experiment. Another paramount factor to consider is the necessity of accurate statistical significance to properly control the proportion of false discoveries among the identified microorganisms, and antimicrobial-resistant proteins, and to provide robust biomass estimation. Recently, we have developed Microorganism Classification and Identification (MiCId) workflow that assigns accurate statistical significance to identified microorganisms, antimicrobial-resistant proteins, and biomass estimation. MiCId's workflow is also computationally efficient, taking about 6–17 minutes to process a tandem mass-spectrometry (MS/MS) experiment using computer resources that are available in most laptop and desktop computers, making it a portable workflow. To make data analysis accessible to a broader range of users, beyond users familiar with the Linux environment, we have developed a graphical user interface (GUI) for MiCId's workflow. The GUI brings to users all the functionality of MiCId's workflow in a friendly interface along with tools for data analysis, visualization, and to export results.

We address the problem of rooting an unrooted species tree given a set of unrooted gene trees, under the assumption that gene trees evolve within the model species tree under the multispecies coalescent (MSC) model. Quintet Rooting (QR) is a polynomial time algorithm that was recently proposed for this problem, which is based on the theory developed by Allman, Degnan, and Rhodes that proves the identifiability of rooted 5-taxon trees from unrooted gene trees under the MSC. However, although QR had good accuracy in simulations, its statistical consistency was left as an open problem. We present QR-STAR, a variant of QR with an additional step and a different cost function, and prove that it is statistically consistent under the MSC. Moreover, we derive sample complexity bounds for QR-STAR and show that a particular variant of it based on "short quintets" has polynomial sample complexity. Finally, our simulation study under a variety of model conditions shows that QR-STAR matches or improves on the accuracy of QR. QR-STAR is available in open-source form on github.

We present a framework called the Reasoning Engine, which implements Satisfiability Modulo Theories (SMT)-based methods within a unified computational environment to address diverse biological analysis problems. The Reasoning Engine was used to reproduce results from key scientific studies, as well as supporting new research in stem cell biology. The framework utilizes an intermediate language for encoding partially specified discrete dynamical systems, which bridges the gap between high-level domain-specific languages and low-level SMT solvers. We provide this framework as open source together with various biological case studies, illustrating the synthesis, enumeration, optimization, and reasoning over models consistent with experimental observations to reveal novel biological insights.

Protein complexes are the foundation of all cellular activities, and accurately identifying them is crucial for studying cellular systems. The efficient discovery of protein complexes is a focus of research in the field of bioinformatics. Most existing methods for protein complex identification are based on the structure of the protein-protein interaction (PPI) network, whereas some methods attempt to integrate biological information to enhance the features of the protein network for complex identification. Existing protein complex identification methods are unable to fully integrate network topology information and biological attribute information. Most of these methods are based on homogeneous networks and cannot distinguish the importance of different attributes and protein nodes. To address these issues, a GO attribute Heterogeneous Attention network Embedding (GHAE) method based on heterogeneous protein information networks is proposed. First, GHAE incorporates Gene Ontology (GO) information into the PPI network, constructing a heterogeneous protein information network. Then, GHAE uses a dual attention mechanism and heterogeneous graph convolutional representation learning method to learn protein features and to identify protein complexes. The experimental results show that building heterogeneous protein information networks can fully integrate valuable biological information. The heterogeneous graph embedding learning method can simultaneously mine the features of protein and GO attributes, thereby improving the performance of protein complex identification.

International Classification of Diseases (ICD) serves as the foundation for generating comparable global disease statistics across regions and over time. The process of ICD coding involves assigning codes to diseases based on clinical notes, which can describe a patient's condition in a standard way. However, this process is complicated by the vast number of codes and the intricate taxonomy of ICD codes, which are hierarchically organized into various levels, including chapter, category, subcategory, and its subdivisions. Many existing studies focus solely on predicting subcategory codes, ignoring the hierarchical relationships among codes. To address this limitation, we propose a multitask learning model that trains multiple classifiers for different code levels, while also capturing the relations between coarser and finer-grained labels through a reinforcement mechanism. Our approach is evaluated on both English and Chinese benchmark dataset, and we demonstrate that our method achieves competitive performance with baseline models, particularly in terms of macro-F1 results. These findings suggest that our approach effectively leverages the hierarchical structure of ICD codes to improve disease code prediction accuracy. Analysis of attention mechanism shows that multigranularity attention of our model captures crucial feature of input text on different granularity levels, which can provide reasonable explanations for the prediction results.

With the rapid spread of COVID-19 worldwide, viral genomic data are available in the order of millions of sequences on public databases such as GISAID. This Big Data creates a unique opportunity for analysis toward the research of effective vaccine development for current pandemics, and avoiding or mitigating future pandemics. One piece of information that comes with every such viral sequence is the geographical location where it was collected-the patterns found between viral variants and geographical location surely being an important part of this analysis. One major challenge that researchers face is processing such huge, highly dimensional data to obtain useful insights as quickly as possible. Most of the existing methods face scalability issues when dealing with the magnitude of such data. In this article, we propose an approach that first computes a numerical representation of the spike protein sequence of SARS-CoV-2 using k-mers (substrings) and then uses several machine learning models to classify the sequences based on geographical location. We show that our proposed model significantly outperforms the baselines. We also show the importance of different amino acids in the spike sequences by computing the information gain corresponding to the true class labels.

A major challenge in molecular systems biology is to understand how proteins work to transmit external signals to changes in gene expression. Computationally reconstructing these signaling pathways from protein interaction networks can help understand what is missing from existing pathway databases. We formulate a new pathway reconstruction problem, one that iteratively grows directed acyclic graphs (DAGs) from a set of starting proteins in a protein interaction network. We present an algorithm that provably returns the optimal DAGs for two different cost functions and evaluate the pathway reconstructions when applied to six diverse signaling pathways from the NetPath database. The optimal DAGs outperform an existing k-shortest paths method for pathway reconstruction, and the new reconstructions are enriched for different biological processes. Growing DAGs is a promising step toward reconstructing pathways that provably optimize a specific cost function.

Genome-scale constraint-based metabolic networks play an important role in the simulation of growth-coupled production, which means that cell growth and target metabolite production are simultaneously achieved. For growth-coupled production, a minimal reaction-network-based design is known to be effective. However, the obtained reaction networks often fail to be realized by gene deletions due to conflicts with gene-protein-reaction (GPR) relations. Here, we developed gDel_minRN that determines gene deletion strategies using mixed-integer linear programming to achieve growth-coupled production by repressing the maximum number of reactions via GPR relations. The results of computational experiments showed that gDel_minRN could determine the core parts, which include only 30% to 55% of whole genes, for stoichiometrically feasible growth-coupled production for many target metabolites, which include useful vitamins such as biotin (vitamin B7), riboflavin (vitamin B2), and pantothenate (vitamin B5). Since gDel_minRN calculates a constraint-based model of the minimum number of gene-associated reactions without conflict with GPR relations, it helps biological analysis of the core parts essential for growth-coupled production for each target metabolite. The source codes, implemented in MATLAB using CPLEX and COBRA Toolbox, are available on https://github.com/MetNetComp/gDel-minRN.