WellGreen-i Co., Ltd.

• Routine literature reviews are a standard part of the research process, aimed at summarizing and synthesizing existing knowledge.
• The amount of information available in journals and papers accessible to researchers is enormous.
• As academic fields and journals become increasingly specialized, papers outside one’s area of expertise can be difficult to read and understand, even if they contain information of interest.
• The time, human resources, and funding available for literature reviews are limited.

Notes on Automatically Collected and Analyzed Information in This Contracted Service

• The information on websites and documents handled by this service consists of text data that is authorized for analysis.
• Key conclusions from research projects are presented in the abstracts of papers. Extracting and summarizing sentences containing key term(s) from a large number of papers enables the collection of knowledge-based information and the derivation of various statistical indices from big data.
• Freely accessible full-text papers can also be included in the analysis.
• When the full texts of papers are available for analysis, they can also be used to collect knowledge-based information, produce summaries, and perform statistical processing.

WGI has developed AI text-mining technology, known as A2K (Article-to-Knowledge) Technology, which enables automated, high-precision, large-scale literature reviews to collect key information for each project. Using this technology, WGI offers services that collect and summarize key information from big data.

• key terms can be freely determined according to project focus areas (e.g., viral resistance, breast cancer suppression genes, unsaturated fatty acid content, flowering control).
See more...
Any number of key terms can be used in a single A2K analysis.
English sentences containing key term(s) are comprehensively collected and automatically analyzed for their basic structures (e.g., SVO, SVC), enabling the rapid and highly accurate generation of summarized information about the key terms.
By mechanically processing a large volume of available English texts (including online information and electronic files) in bulk, it is possible to comprehensively and efficiently collect and compile the key information.
The result of the A2K analysis is presented in the A2K Description format, which provides an easy-to-understand summary of information about key terms.

In A2K analysis, it can be difficult to comprehensively aggregate gene information when using gene names as key terms.

• Different names (synonyms) may be assigned to the same gene, and the terminology used can vary across publications. Consequently, in A2K analysis, it can be challenging to comprehensively aggregate gene information when using gene names as key terms.
• Therefore, when aiming to collect functional information on a gene, it is necessary to include all known synonyms as key terms.
• However, since new synonyms continue to emerge with the publication of new papers, listing all of them is practically impossible.

LA2K technology incorporates an AI-driven named-entity recognition (NER) method that automatically identifies gene names in text.

• Using LA2K technology, gene information described in text can be comprehensively and automatically extracted.
• LA2K analysis of peer-reviewed papers provides knowledge-based information on genes in the text.
• Since LA2K technology is computer-processed, certain gene names with specific patterns (e.g., those consisting solely of numbers) are difficult to automatically recognize.
• The accuracy of gene name recognition by LA2K technology is expected to improve further through the accumulation of analytical data and refinement of its algorithms.

LA2K technology collects documents related to the key term(s) and extracts sentences containing the key term(s) and/or gene name(s). In this process, the gene names that appear in the document are automatically recognized by the LA2K technology. The relevant information from these sentences is then summarized and presented in the A2K Description format.

An example of analysis using the LA2K technique is presented in the following figure. LA2K technology extracts key information related not only to the key term(s) but also to gene(s). In this example, the A2K Descriptions of the automatically recognized gene names in the document are shown.

An example of LA2K analysis using key terms:breast cancer, gene expression profiling, microarray analysis, and estrogen receptor alpha.

• Gene names automatically recognized in text by LA2K analysis are shown in the column "Gene name" in the table.
• The key information on the genes is summarized and presented in the A2K Description.
• Although A2K Descriptions related to the key terms are also obtained, they are not shown in the table.

In this way, by utilizing LA2K technology, the computer can recognize automatically and correctly gene names in the text just like manual review by a researcher. Owing to the powerful text analysis capability of LA2K technology, key information related to genes can be extracted from documents even without prior knowledge of gene names by either the computer or the user.

• The entire process—from literature search and collection to the extraction of knowledge-based information (key information derived from the summarization of research findings) related to the key terms and genes—is automatically performed.
• Users of LA2K technology can easily analyze large volumes of literature simply by providing key terms of interest.
• Knowledge-based information can be quickly and precisely obtained even when a large volume of literature is analyzed.
• While researchers may have difficulties comprehending the contents of articles outside their specialty area, LA2K technology can handle papers from any field.
• Researchers are liberated from the substantial time, effort, and cost typically required for daily literature surveys.
• By considerably shortening the time required for literature surveys, researchers can focus on advancing their projects, thereby enabling shorter timeframes for achieving goals and enhancing the quality of results.

Integrating of LA2K analysis results with omics data maximize the information for elucidating the molecular mechanisms of various biological processes. Information on transcription factors and cis-elements, gene expression data, and homolog information (gene families) within and across species are valuable as the omics data for the integration. Such integration of multifaceted information facilitates the achievement of research goals, including gene identification.

• Simply by providing key terms of interest, the entire process—from literature search to knowledge-based information compilation—can be executed seamlessly and automatically within a short time.
• A vast amount of text information (numerous documents) related to the key terms can be collectively gathered and summarized.
• Experts can significantly reduce the amount of text they need to read and comprehend on their own.
• While manual review becomes difficult when papers fall outside a researcher’s specialty, LA2K technology is capable of handling papers from any field―medicine, pharmacy, basic biology, agriculture, bioinformatics, patents, and beyond.
• There is no limit to the number of key terms used for A2K analysis.
• A2K analysis is conducted on English texts.
• By utilizing automatic translation tools, Japanese texts and others can also be analyzed (note that the accuracy of translation affects the precision of summarization and extraction of key information).