National Cancer Center released the application and challenges of big data in cancer research and clinical conversion

#cancer#

Humans have a long history of the research of cancer mechanism and therapy. The name of "tumor" has appeared in Yin Ruine Oracle 3600 years ago. With the maturity of molecular biology technology in the last century, the mechanism of the development of many cancers was discovered in the study of some gene molecules, and some anticancer drugs with certain effects were also born. Observation of clinical experience since ancient times and the study of molecular mechanisms since the last century have been the main driving force for scientific progress of cancer. In recent years, the breakthroughs of Qualcomm's technology and the rapid development of computer science have begun to guide cancer research, clinical diagnosis, and new drug research and development have entered a model of big data and artificial intelligence -driven. Many new mechanism discovery and clinical applications were born under big data catalysis. But at the same time, many of the effectiveness of artificial intelligence and data models have been exaggerated. Many news media reports are often unsatisfactory in practical applications.

So as a cancer scientific research and clinical worker, how can we effectively use big data resources while avoiding the limitations and deviations that the current data science may bring?

On September 5, 2022, Dr. Jiang Peng and Eytan RUPPIN in the National Cancer Center of the United States published a review article on Nature Reviews Cancer Big Data in Basic and Translational Cancer Research. Application and challenges.

Big data itself is a concept related to the times. The deep blue computing power to defeat Casparov in 1997 is far less than most smartphones in 2022. Therefore, the author of the paper first gives big data a dynamic definition: large computer clusters are needed in contemporary research to store and processes, and new discoveries on some important issues can be promoted. Under this definition, this summary first lists important data types, large -scale data resources, and data interactive analysis platforms related to cancer research. These resources are not only useful for computing professionals, but also many interactive network platforms can also help researchers who do not programming use large -scale data resources to study new issues. Then, we discussed several common strategies in data -driven research, such as cross -mode integration, cross -queue integration, and knowledge transfer. These data resources and analysis strategies have catalyzed a lot of discoveries, so that we re -understand the process of cancer development, and also promoted a group of emerging sectors of diagnostic products or industrial circles that have been put into use in clinical practice.

In addition to enumerating the current achievements and opportunities, this review also focuses on many systemic limitations of cancer data science. Compared with other computer disciplines, such as machine vision, many basic problems of cancer big data science need to be further resolved. For example, the amount of data is very limited, resulting in unreliable artificial intelligence models. The interpretation and marking of data lack of unified standards, resulting in a large number of published resources cannot be uniformly used. Moreover, the data sharing mechanism and culture have not been established in the field of cancer data science, which has led to a lot of valuable data sets that cost a lot of resources, even after the publication is not truly public. To solve these problems, it is not only a scientific issue. It is necessary to involve the formulation and implementation of many fund issuance and use policies, third -party regulatory mechanisms, and flexible open source sharing models.

Dr. Jiang Peng and Dr. Eytan Ruppin are the joint communication authors of the article. Jiang Peng and the first author of this article. Sanju Sinha, Kenneth Aldape, Sridhar Hannenhalli, Cenk Sahinalp also participated in writing.

Original link:

https://www.nature.com/articleS/S41568-022-00502-0

PDF can be downloaded from this:

https://t.co/neapgur6xe

Want to know more exciting content, come and pay attention to BIOART biological art

- END -