Availability of annotated text data

Large annotated corpora used in the development and training of general purpose text mining methods (e.g., sets of movie dialogue,^[3] product reviews,^[4] or Wikipedia article text) are not specific for biomedical language. While they may provide evidence of general text properties such as parts of speech, they rarely contain concepts of interest to biologists or clinicians. Development of new methods to identify features specific to biomedical documents therefore requires assembly of specialized corpora.^[5] Resources designed to aid in building new biomedical text mining methods have been developed through the Informatics for Integrating Biology and the Bedside (i2b2) challenges^[6][7][8] and biomedical informatics researchers.^[9][10] Text mining researchers frequently combine these corpora with the controlled vocabularies and ontologies available through the National Library of Medicine's Unified Medical Language System (UMLS) and Medical Subject Headings (MeSH).

Machine learning-based methods often require very large data sets as training data to build useful models.^[11] Manual annotation of large text corpora is not realistically possible. Training data may therefore be products of weak supervision^[12][13] or purely statistical methods.

Data structure variation

Like other text documents, biomedical documents contain unstructured data.^[14] Research publications follow different formats, contain different types of information, and are interspersed with figures, tables, and other non-text content. Both unstructured text and semi-structured document elements, such as tables, may contain important information that should be text mined.^[15] Clinical documents may vary in structure and language between departments and locations. Other types of biomedical text, such as drug labels,^[16] may follow general structural guidelines but lack further details.

Uncertainty

Biomedical literature contains statements about observations that may not be statements of fact. This text may express uncertainty or skepticism about claims. Without specific adaptations, text mining approaches designed to identify claims within text may mis-characterize these "hedged" statements as facts.^[17]

Supporting clinical needs

Biomedical text mining applications developed for clinical use should ideally reflect the needs and demands of clinicians.^[5] This is a concern in environments where clinical decision support is expected to be informative and accurate. A comprehensive overview of the development and uptake of NLP methods applied to free-text clinical notes related to chronic diseases is presented in.^[18]

Interoperability with clinical systems

New text mining systems must work with existing standards, electronic medical records, and databases.^[5] Methods for interfacing with clinical systems such as LOINC have been developed^[19] but require extensive organizational effort to implement and maintain.^[20][21]

Patient privacy

Text mining systems operating with private medical data must respect its security and ensure it is rendered anonymous where appropriate.^[22][23][24]

Named entity recognition

Developments in biomedical text mining have incorporated identification of biological entities with named entity recognition, or NER. Names and identifiers for biomolecules such as proteins and genes,^[25] chemical compounds and drugs,^[26] and disease names^[27] have all been used as entities. Most entity recognition methods are supported by pre-defined linguistic features or vocabularies, though methods incorporating deep learning and word embeddings have also been successful at biomedical NER.^[28][29]

Document classification and clustering

Biomedical documents may be classified or clustered based on their contents and topics. In classification, document categories are specified manually,^[30] while in clustering, documents form algorithm-dependent, distinct groups.^[31] These two tasks are representative of supervised and unsupervised methods, respectively, yet the goal of both is to produce subsets of documents based on their distinguishing features. Methods for biomedical document clustering have relied upon k-means clustering.^[31]

Relationship discovery

Biomedical documents describe connections between concepts, whether they are interactions between biomolecules, events occurring subsequently over time (i.e., temporal relationships), or causal relationships. Text mining methods may perform relation discovery to identify these connections, often in concert with named entity recognition.^[32]

Hedge cue detection

The challenge of identifying uncertain or "hedged" statements has been addressed through hedge cue detection in biomedical literature.^[17]

Claim detection

Multiple researchers have developed methods to identify specific scientific claims from literature.^[33][34] In practice, this process involves both isolating phrases and sentences denoting the core arguments made by the authors of a document (a process known as argument mining, employing tools used in fields such as political science) and comparing claims to find potential contradictions between them.^[34]

Information extraction

Information extraction, or IE, is the process of automatically identifying structured information from unstructured or partially structured text. IE processes can involve several or all of the above activities, including named entity recognition, relationship discovery, and document classification, with the overall goal of translating text to a more structured form, such as the contents of a template or knowledge base. In the biomedical domain, IE is used to generate links between concepts described in text, such as gene A inhibits gene B and gene C is involved in disease G.^[35] Biomedical knowledge bases containing this type of information are generally products of extensive manual curation, so replacement of manual efforts with automated methods remains a compelling area of research.^[36][37]

Information retrieval and question answering

Biomedical text mining supports applications for identifying documents and concepts matching search queries. Search engines such as PubMed search allow users to query literature databases with words or phrases present in document contents, metadata, or indices such as MeSH. Similar approaches may be used for medical literature retrieval. For more fine-grained results, some applications permit users to search with natural language queries and identify specific biomedical relationships.^[38]

On 16 March 2020, the National Library of Medicine and others launched the COVID-19 Open Research Dataset (CORD-19) to enable text mining of the current literature on the novel virus. The dataset is hosted by the Semantic Scholar project^[39] of the Allen Institute for AI.^[40] Other participants include Google, Microsoft Research, the Center for Security and Emerging Technology, and the Chan Zuckerberg Initiative.^[41]

Corpora

The following table lists a selection of biomedical text corpora and their contents. These items include annotated corpora, sources of biomedical research literature, and resources frequently used as vocabulary and/or ontology references, such as MeSH. Items marked "Yes" under "Freely Available" can be downloaded from a publicly accessible location.

Word embeddings

Several groups have developed sets of biomedical vocabulary mapped to vectors of real numbers, known as word vectors or word embeddings. Sources of pre-trained embeddings specific for biomedical vocabulary are listed in the table below. The majority are results of the word2vec model developed by Mikolov et al^[86] or variants of word2vec.

Gene cluster identification

Methods for determining the association of gene clusters obtained by microarray experiments with the biological context provided by the corresponding literature have been developed.^[97]

Protein interactions

Automatic extraction of protein interactions^[98] and associations of proteins to functional concepts (e.g. gene ontology terms) has been explored.^{[citation needed]} The search engine PIE was developed to identify and return protein-protein interaction mentions from MEDLINE-indexed articles.^[99] The extraction of kinetic parameters from text or the subcellular location of proteins have also been addressed by information extraction and text mining technology.^{[citation needed]}

Gene-disease associations

Computational gene prioritization is an essential step in understanding the genetic basis of diseases, particularly within genetic linkage analysis. Text mining and other computational tools extract relevant information, including gene-disease associations, among others, from numerous data sources, then apply different ranking algorithms to prioritize the genes based on their relevance to the specific disease.^[100] Text mining and gene prioritization allow researchers to focus their efforts on the most promising candidates for further research.

Computational tools for gene prioritization continue to be developed and analyzed. One group studied the performance of various text-mining techniques for disease gene prioritization. They investigated different domain vocabularies, text representation schemes, and ranking algorithms in order to find the best approach for identifying disease-causing genes to establish a benchmark.^[101]

Gene-trait associations

An agricultural genomics group identified genes related to bovine reproductive traits using text mining, among other approaches.^[102]

Search engines

Search engines designed to retrieve biomedical literature relevant to a user-provided query frequently rely upon text mining approaches. Publicly available tools specific for research literature include PubMed search, Europe PubMed Central search, GeneView,^[104] and APSE^[105] Similarly, search engines and indexing systems specific for biomedical data have been developed, including DataMed^[106] and OmicsDI.^[107]

Some search engines, such as Essie,^[108] OncoSearch,^[109] PubGene,^[110][111] and GoPubMed^[112] were previously public but have since been discontinued, rendered obsolete, or integrated into commercial products.

Medical record analysis systems

Electronic medical records (EMRs) and electronic health records (EHRs) are collected by clinical staff in the course of diagnosis and treatment. Though these records generally include structured components with predictable formats and data types, the remainder of the reports are often free-text and difficult to search, leading to challenges with patient care.^[113] Numerous complete systems and tools have been developed to analyse these free-text portions.^[114] The MedLEE system was originally developed for analysis of chest radiology reports but later extended to other report topics.^[115] The clinical Text Analysis and Knowledge Extraction System, or cTAKES, annotates clinical text using a dictionary of concepts.^[116] The CLAMP system offers similar functionality with a user-friendly interface.^[117]

Frameworks

Computational frameworks have been developed to rapidly build tools for biomedical text mining tasks. SwellShark^[118] is a framework for biomedical NER that requires no human-labeled data but does make use of resources for weak supervision (e.g., UMLS semantic types). The SparkText framework^[119] uses Apache Spark data streaming, a NoSQL database, and basic machine learning methods to build predictive models from scientific articles.

APIs

Some biomedical text mining and natural language processing tools are available through application programming interfaces, or APIs. NOBLE Coder performs concept recognition through an API.^[120]