As the availability of big biomedical data advances, there is a growing need of university students trained professionally on analyzing these data and correctly interpreting their results. We propose here a study plan for a master's degree course on biomedical data science, by describing our experience during the last academic year. In our university course, we explained how to find an open biomedical dataset, how to correctly clean it and how to prepare it for a computational statistics or machine learning phase. By doing so, we introduce common health data science terms and explained how to avoid common mistakes in the process. Moreover, we clarified how to perform an exploratory data analysis (EDA) and how to reasonably interpret its results. We also described how to properly execute a supervised or unsupervised machine learning analysis, and now to understand and interpret its outcomes. Eventually, we explained how to validate the findings obtained. We illustrated all these steps in the context of open science principles, by suggesting to the students to use only open source programming languages (R or Python in particular), open biomedical data (if available), and open access scientific articles (if possible). We believe our teaching proposal can be useful and of interest for anyone wanting to start to prepare a course on biomedical data science.

A recent study by Faye Orcales and colleagues proposes a teaching curriculum on supervised machine learning applied to genomics data aimed at predicting antibiotic resistance. The article describes a traditional machine learning pipeline step-by-step in a way that is accessible to anyone, including novices. However, the authors provide a misleading piece of advice in the "Evaluating model performance" section, where they recommend that readers use accuracy and the F1 score for binary classification. We write this short formal comment on that article to reaffirm and explain why accuracy and the F1 score should be avoided in the evaluation of binary classification and why the Matthews correlation coefficient (MCC) should be employed instead. We also take this opportunity to warn readers about the dangers of k-fold cross-validation, which is suggested as a standard method for dividing data into training set and test set, but has several flaws and pitfalls.

Machine learning has become a powerful tool for computational analysis in the biomedical sciences, with its effectiveness significantly enhanced by integrating domain-specific knowledge. This integration has give rise to informed machine learning, in contrast to studies that lack domain knowledge and treat all variables equally (uninformed machine learning). While the application of informed machine learning to bioinformatics and health informatics datasets has become more seamless, the likelihood of errors has also increased. To address this drawback, we present eight guidelines outlining best practices for employing informed machine learning methods in biomedical sciences. These quick tips offer recommendations on various aspects of informed machine learning analysis, aiming to assist researchers in generating more robust, explainable, and dependable results. Even if we originally crafted these eight simple suggestions for novices, we believe they are deemed relevant for expert computational researchers as well.

As machine learning (ML) becomes increasingly central to biomedical research, the need for trustworthy models is more pressing than ever. In this paper, we present nine concise and actionable tips to help researchers build ML systems that are technically sound but ethically responsible, and contextually appropriate for biomedical applications. These tips address the multifaceted nature of trustworthiness, emphasizing the importance of considering all potential consequences, recognizing the limitations of current methods, taking into account the needs of all involved stakeholders, and following open science practices. We discuss technical, ethical, and domain-specific challenges, offering guidance on how to define trustworthiness and how to mitigate sources of untrustworthiness. By embedding trustworthiness into every stage of the ML pipeline - from research design to deployment - these recommendations aim to support both novice and experienced practitioners in creating ML systems that can be relied upon in biomedical science.

Over the past two decades, extensive studies, particularly in cancer analysis through large datasets like The Cancer Genome Atlas (TCGA), have aimed at improving patient therapies and precision medicine. However, limited overlap and inconsistencies among gene signatures across different cohorts pose challenges. The dynamic nature of the transcriptome, encompassing diverse RNA species and functional complexities at gene and isoform levels, introduces intricacies, and current gene signatures face reproducibility issues due to the unique transcriptomic landscape of each patient. In this context, discrepancies arising from diverse sequencing technologies, data analysis algorithms, and software tools further hinder consistency. While careful experimental design, analytical strategies, and standardized protocols could enhance reproducibility, future prospects lie in multiomics data integration, machine learning techniques, open science practices, and collaborative efforts. Standardized metrics, quality control measures, and advancements in single-cell RNA-seq will contribute to unbiased gene signature identification. In this perspective article, we outline some thoughts and insights addressing challenges, standardized practices, and advanced methodologies enhancing the reliability of gene signatures in disease transcriptomic research.

Data are the most important elements of bioinformatics: Computational analysis of bioinformatics data, in fact, can help researchers infer new knowledge about biology, chemistry, biophysics, and sometimes even medicine, influencing treatments and therapies for patients. Bioinformatics and high-throughput biological data coming from different sources can even be more helpful, because each of these different data chunks can provide alternative, complementary information about a specific biological phenomenon, similar to multiple photos of the same subject taken from different angles. In this context, the integration of bioinformatics and high-throughput biological data gets a pivotal role in running a successful bioinformatics study. In the last decades, data originating from proteomics, metabolomics, metagenomics, phenomics, transcriptomics, and epigenomics have been labelled -omics data, as a unique name to refer to them, and the integration of these omics data has gained importance in all biological areas. Even if this omics data integration is useful and relevant, due to its heterogeneity, it is not uncommon to make mistakes during the integration phases. We therefore decided to present these ten quick tips to perform an omics data integration correctly, avoiding common mistakes we experienced or noticed in published studies in the past. Even if we designed our ten guidelines for beginners, by using a simple language that (we hope) can be understood by anyone, we believe our ten recommendations should be taken into account by all the bioinformaticians performing omics data integration, including experts.

Some scientific studies involve huge amounts of bioinformatics data that cannot be analyzed on personal computers usually employed by researchers for day-to-day activities but rather necessitate effective computational infrastructures that can work in a distributed way. For this purpose, distributed computing systems have become useful tools to analyze large amounts of bioinformatics data and to generate relevant results on virtual environments, where software can be executed for hours or even days without affecting the personal computer or laptop of a researcher. Even if distributed computing resources have become pivotal in multiple bioinformatics laboratories, often researchers and students use them in the wrong ways, making mistakes that can cause the distributed computers to underperform or that can even generate wrong outcomes. In this context, we present here ten quick tips for the usage of Apache Spark distributed computing systems for bioinformatics analyses: ten simple guidelines that, if taken into account, can help users avoid common mistakes and can help them run their bioinformatics analyses smoothly. Even if we designed our recommendations for beginners and students, they should be followed by experts too. We think our quick tips can help anyone make use of Apache Spark distributed computing systems more efficiently and ultimately help generate better, more reliable scientific results.

Medical imaging is a great asset for modern medicine, since it allows physicians to spatially interrogate a disease site, resulting in precise intervention for diagnosis and treatment, and to observe particular aspect of patients' conditions that otherwise would not be noticeable. Computational analysis of medical images, moreover, can allow the discovery of disease patterns and correlations among cohorts of patients with the same disease, thus suggesting common causes or providing useful information for better therapies and cures. Machine learning and deep learning applied to medical images, in particular, have produced new, unprecedented results that can pave the way to advanced frontiers of medical discoveries. While computational analysis of medical images has become easier, however, the possibility to make mistakes or generate inflated or misleading results has become easier, too, hindering reproducibility and deployment. In this article, we provide ten quick tips to perform computational analysis of medical images avoiding common mistakes and pitfalls that we noticed in multiple studies in the past. We believe our ten guidelines, if taken into practice, can help the computational-medical imaging community to perform better scientific research that eventually can have a positive impact on the lives of patients worldwide.

Fuzzy logic is useful tool to describe and represent biological or medical scenarios, where often states and outcomes are not only completely true or completely false, but rather partially true or partially false. Despite its usefulness and spread, fuzzy logic modeling might easily be done in the wrong way, especially by beginners and unexperienced researchers, who might overlook some important aspects or might make common mistakes. Malpractices and pitfalls, in turn, can lead to wrong or overoptimistic, inflated results, with negative consequences to the biomedical research community trying to comprehend a particular phenomenon, or even to patients suffering from the investigated disease. To avoid common mistakes, we present here a list of quick tips for fuzzy logic modeling any biomedical scenario: some guidelines which should be taken into account by any fuzzy logic practitioner, including experts. We believe our best practices can have a strong impact in the scientific community, allowing researchers who follow them to obtain better, more reliable results and outcomes in biomedical contexts.

SUMMARY: Having multiple datasets is a key aspect of robust bioinformatics analyses, because it allows researchers to find possible confirmation of the discoveries made on multiple cohorts. For this purpose, Gene Expression Omnibus (GEO) can be a useful database, since it provides hundreds of thousands of microarray gene expression datasets freely available for download and usage. Despite this large availability, collecting prognostic datasets of a specific cancer type from GEO can be a long, time-consuming and energy-consuming activity for any bioinformatician, who needs to execute it manually by first performing a search on the GEO website and then by checking all the datasets found one by one. To solve this problem, we present here geoCancerPrognosticDatasetsRetriever, a Perl 5 application which reads a cancer type and a list of microarray platforms, searches for prognostic gene expression datasets of that cancer type and based on those platforms available on GEO, and returns the GEO accession codes of those datasets, if found. Our bioinformatics tool can easily generate in a few minutes a list of cancer prognostic datasets that otherwise would require numerous hours of manual work to any bioinformatician. geoCancerPrognosticDatasetsRetriever can handily retrieve multiple prognostic datasets of gene expression of any cancer type, laying the foundations for numerous bioinformatics studies and meta-analyses that can have a strong impact on oncology research. AVAILABILITY AND IMPLEMENTATION: geoCancerPrognosticDatasetsRetriever is freely available under the GPLv2 license on the Comprehensive Perl Archive Network (CPAN) at https://metacpan.org/pod/App::geoCancerPrognosticDatasetsRetriever and on GitHub at https://github.com/AbbasAlameer/geoCancerPrognosticDatasetsRetriever. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Pathway enrichment analysis (PEA) is a computational biology method that identifies biological functions that are overrepresented in a group of genes more than would be expected by chance and ranks these functions by relevance. The relative abundance of genes pertinent to specific pathways is measured through statistical methods, and associated functional pathways are retrieved from online bioinformatics databases. In the last decade, along with the spread of the internet, higher availability of computational resources made PEA software tools easy to access and to use for bioinformatics practitioners worldwide. Although it became easier to use these tools, it also became easier to make mistakes that could generate inflated or misleading results, especially for beginners and inexperienced computational biologists. With this article, we propose nine quick tips to avoid common mistakes and to out a complete, sound, thorough PEA, which can produce relevant and robust results. We describe our nine guidelines in a simple way, so that they can be understood and used by anyone, including students and beginners. Some tips explain what to do before starting a PEA, others are suggestions of how to correctly generate meaningful results, and some final guidelines indicate some useful steps to properly interpret PEA results. Our nine tips can help users perform better pathway enrichment analyses and eventually contribute to a better understanding of current biology.

Special sessions are important parts of scientific meetings and conferences: They gather together researchers and students interested in a specific topic and can strongly contribute to the success of the conference itself. Moreover, they can be the first step for trainees and students to the organization of a scientific event. Organizing a special session, however, can be uneasy for beginners and students. Here, we provide ten simple rules to follow to organize a special session at a scientific conference.

Applying computational statistics or machine learning methods to data is a key component of many scientific studies, in any field, but alone might not be sufficient to generate robust and reliable outcomes and results. Before applying any discovery method, preprocessing steps are necessary to prepare the data to the computational analysis. In this framework, data cleaning and feature engineering are key pillars of any scientific study involving data analysis and that should be adequately designed and performed since the first phases of the project. We call "feature" a variable describing a particular trait of a person or an observation, recorded usually as a column in a dataset. Even if pivotal, these data cleaning and feature engineering steps sometimes are done poorly or inefficiently, especially by beginners and unexperienced researchers. For this reason, we propose here our quick tips for data cleaning and feature engineering on how to carry out these important preprocessing steps correctly avoiding common mistakes and pitfalls. Although we designed these guidelines with bioinformatics and health informatics scenarios in mind, we believe they can more in general be applied to any scientific area. We therefore target these guidelines to any researcher or practitioners wanting to perform data cleaning or feature engineering. We believe our simple recommendations can help researchers and scholars perform better computational analyses that can lead, in turn, to more solid outcomes and more reliable discoveries.

Myocardial infarctions and heart failure are the cause of more than 17 million deaths annually worldwide. ST-segment elevation myocardial infarctions (STEMI) require timely treatment, because delays of minutes have serious clinical impacts. Machine learning can provide alternative ways to predict heart failure and identify genes involved in heart failure. For these scopes, we applied a Random Forests classifier enhanced with feature elimination to microarray gene expression of 111 patients diagnosed with STEMI, and measured the classification performance through standard metrics such as the Matthews correlation coefficient (MCC) and area under the receiver operating characteristic curve (ROC AUC). Afterwards, we used the same approach to rank all genes by importance, and to detect the genes more strongly associated with heart failure. We validated this ranking by literature review and gene set enrichment analysis. Our classifier employed to predict heart failure achieved MCC = +0.87 and ROC AUC = 0.918, and our analysis identified KLHL22, WDR11, OR4Q3, GPATCH3, and FAH as top five protein-coding genes related to heart failure. Our results confirm the effectiveness of machine learning feature elimination in predicting heart failure from gene expression, and the top genes found by our approach will be able to help biologists and cardiologists further our understanding of heart failure.

Biomolecular controlled annotations have become pivotal in computational biology, because they allow scientists to analyze large amounts of biological data to better understand test results, and to infer new knowledge. Yet, biomolecular annotation databases are incomplete by definition, like our knowledge of biology, and might contain errors and inconsistent information. In this context, machine-learning algorithms able to predict and prioritize new annotations are both effective and efficient, especially if compared with time-consuming trials of biological validation. To limit the possibility that these techniques predict obvious and trivial high-level features, and to help prioritize their results, we introduce a new element that can improve accuracy and relevance of the results of an annotation prediction and prioritization pipeline. We propose a novelty indicator able to state the level of "originality" of the annotations predicted for a specific gene to Gene Ontology (GO) terms. This indicator, joint with our previously introduced prediction steps, helps by prioritizing the most novel interesting annotations predicted. We performed an accurate biological functional analysis of the prioritized annotations predicted with high accuracy by our indicator and previously proposed methods. The relevance of our biological findings proves effectiveness and trustworthiness of our indicator and of its prioritization of predicted annotations.

Genes and their protein products are essential molecular units of a living organism. The knowledge of their functions is key for the understanding of physiological and pathological biological processes, as well as in the development of new drugs and therapies. The association of a gene or protein with its functions, described by controlled terms of biomolecular terminologies or ontologies, is named gene functional annotation. Very many and valuable gene annotations expressed through terminologies and ontologies are available. Nevertheless, they might include some erroneous information, since only a subset of annotations are reviewed by curators. Furthermore, they are incomplete by definition, given the rapidly evolving pace of biomolecular knowledge. In this scenario, computational methods that are able to quicken the annotation curation process and reliably suggest new annotations are very important. Here, we first propose a computational pipeline that uses different semantic and machine learning methods to predict novel ontology-based gene functional annotations; then, we introduce a new semantic prioritization rule to categorize the predicted annotations by their likelihood of being correct. Our tests and validations proved the effectiveness of our pipeline and prioritization of predicted annotations, by selecting as most likely manifold predicted annotations that were later confirmed.

BACKGROUND: Gene function annotations, which are associations between a gene and a term of a controlled vocabulary describing gene functional features, are of paramount importance in modern biology. Datasets of these annotations, such as the ones provided by the Gene Ontology Consortium, are used to design novel biological experiments and interpret their results. Despite their importance, these sources of information have some known issues. They are incomplete, since biological knowledge is far from being definitive and it rapidly evolves, and some erroneous annotations may be present. Since the curation process of novel annotations is a costly procedure, both in economical and time terms, computational tools that can reliably predict likely annotations, and thus quicken the discovery of new gene annotations, are very useful. METHODS: We used a set of computational algorithms and weighting schemes to infer novel gene annotations from a set of known ones. We used the latent semantic analysis approach, implementing two popular algorithms (Latent Semantic Indexing and Probabilistic Latent Semantic Analysis) and propose a novel method, the Semantic IMproved Latent Semantic Analysis, which adds a clustering step on the set of considered genes. Furthermore, we propose the improvement of these algorithms by weighting the annotations in the input set. RESULTS: We tested our methods and their weighted variants on the Gene Ontology annotation sets of three model organism genes (Bos taurus, Danio rerio and Drosophila melanogaster ). The methods showed their ability in predicting novel gene annotations and the weighting procedures demonstrated to lead to a valuable improvement, although the obtained results vary according to the dimension of the input annotation set and the considered algorithm. CONCLUSIONS: Out of the three considered methods, the Semantic IMproved Latent Semantic Analysis is the one that provides better results. In particular, when coupled with a proper weighting policy, it is able to predict a significant number of novel annotations, demonstrating to actually be a helpful tool in supporting scientists in the curation process of gene functional annotations.

In the computational biology community, machine learning algorithms are key instruments for many applications, including the prediction of gene-functions based upon the available biomolecular annotations. Additionally, they may also be employed to compute similarity between genes or proteins. Here, we describe and discuss a software suite we developed to implement and make publicly available some of such prediction methods and a computational technique based upon Latent Semantic Indexing (LSI), which leverages both inferred and available annotations to search for semantically similar genes. The suite consists of three components. BioAnnotationPredictor is a computational software module to predict new gene-functions based upon Singular Value Decomposition of available annotations. SimilBio is a Web module that leverages annotations available or predicted by BioAnnotationPredictor to discover similarities between genes via LSI. The suite includes also SemSim, a new Web service built upon these modules to allow accessing them programmatically. We integrated SemSim in the Bio Search Computing framework (http://www.bioinformatics.deib. polimi.it/bio-seco/seco/), where users can exploit the Search Computing technology to run multi-topic complex queries on multiple integrated Web services. Accordingly, researchers may obtain ranked answers involving the computation of the functional similarity between genes in support of biomedical knowledge discovery.

Gene function annotations are key elements in biology and bioinformatics. A typical annotation is the association between a gene and a feature term that describes a functional feature of the gene by using a controlled vocabulary term (e.g. a Gene Ontology (GO) feature term). Unfortunately, available annotations contain errors and biologically validated ones are incomplete by definition, since new knowledge is continuously discovered. Thus, computational algorithms which are able to provide ranked lists of predicted new gene annotations are an excellent contribution to the bioinformatics research. Here, we propose two variants of the known Latent Dirichlet Allocation (LDA) algorithm applied to the prediction of gene annotations. LDA is a very efficient machine learning method built on a set of multinomial probability distributions over a set of topics, given a document (a gene, in our case), and on a set of multinomial probability distributions over a set of words (feature terms, in our case), given a topic. In topic modeling, a topic can be considered as a latent meta-category of words, and a document as a mixture of topics. Our two LDA variants use the collapsed Gibbs Sampling method during the training phase, with two distinct initialization approaches to adapt the LDA mathematical model to the biomolecular annotation scenario. Using six outdated datasets of GO annotations of human and brown rat genes, we compared the annotations predicted by our methods to the ones given by the truncated Singular Value Decomposition (tSVD) method previously developed; then, we validated them by using the annotations available in an updated version of the same datasets. Obtained results show the efficiency of our new proposed algorithms.

Truncated Singular Value Decomposition (SVD) has always been a key algorithm in modern machine learning. Scientists and researchers use this applied mathematics method in many fields. Despite a long history and prevalence, the issue of how to choose the best truncation level still remains an open challenge. In this paper, we describe a new algorithm, akin a the discrete optimization method, that relies on the Receiver Operating Characteristics (ROC) Areas Under the Curve (AUCs) computation. We explore a concrete application of the algorithm to a bioinformatics problem, i.e. the prediction of biomolecular annotations. We applied the algorithm to nine different datasets and the obtained results demonstrate the effectiveness of our technique.

Genomic annotations with functional controlled terms, such as the Gene Ontology (GO) ones, are paramount in modern biology. Yet, they are known to be incomplete, since the current biological knowledge is far to be definitive. In this scenario, computational methods that are able to support and quicken the curation of these annotations can be very useful. In a previous work, we discussed the benefits of using the Probabilistic Latent Semantic Analysis algorithm in order to predict novel GO annotations, compared to some Singular Value Decomposition (SVD) based approaches. In this paper, we propose a further enhancement of that method, which aims at weighting the available associations between genes and functional terms before using them as input to the predictive system. The tests that we performed on the annotations of human genes to GO functional terms showed the efficacy of our approach.

Consistency and completeness of biomolecular annotations is a keypoint of correct interpretation of biological experiments. Yet, the associations between genes (or proteins) and features correctly annotated are just some of all the existing ones. As time goes by, they increase in number and become more useful, but they remain incomplete and some of them incorrect. To support and quicken their time-consuming curation procedure and to improve consistence of available annotations, computational methods that are able to supply a ranked list of predicted annotations are hence extremely useful. Starting from a previous work on the automatic prediction of Gene Ontology (GO) annotations based on the Singular Value Decomposition of the annotation matrix, where every matrix element corresponds to the association of a gene with a feature, we propose the use of a modified Probabilistic Latent Semantic Analysis (pLSA) algorithm, named pLSAnorm, to better perform such prediction. pLSA is a statistical technique from the natural language processing field, which has not been used in bioinformatics annotation prediction yet; it takes advantage of the latent information contained in the analyzed data co-occurrences. We proved the effectiveness of the pLSAnorm prediction method by performing k-fold cross-validation of the GO annotations of two organisms, Gallus gallus and Bos taurus. Obtained results demonstrate the efficacy of our approach.

Genomic annotations describing structural and functional features of genes and gene products through controlled terminologies and ontologies are extremely valuable, especially for computational analyses aimed at inferring new biomedical knowledge, which rely on available annotations. Yet, they are incomplete, especially for recently studied genomes, and only some of available annotations represent highly reliable human curated information. In order to help and speedup the time-consuming curation process and improve available annotations, computational methods able to provide prioritized lists of predicted annotations are paramount. Starting from a previous work on automatic prediction of Gene Ontology annotations based on singular value decomposition (SVD) of gene-to-term annotation matrix, here we propose a novel prediction algorithm that incorporates gene clustering based on gene functional similarity computed on Gene Ontology annotations. We tested both prediction methods performing k-fold cross-validation on two organism genomes, Saccharomyces cerevisiae (SGD) and Drosophila melanogaster (FlyBase). Results demonstrate effectiveness of our approach.