Mass spectrometric data analysis apparatus and analysis method

Abstract

An objective of the invention is, when cancer determination is performed by machine learning using mass spectrum data on a sample taken from a subject, to ensure independence from the determination bya tumor marker test performed in parallel and to increase the reliability of mutual diagnosis. The m / z value of ions derived from substances, used in other tests, such as tumor markers is stored in atarget m / z value DB (31), and a spectrum information filtering unit (22) deletes the signal intensity of the m / z value stored in the target m / z value DB (31) from a large number of mass spectra thathave been classified according to the presence or absence of cancer. A learning processing unit (23) obtains the learning result information by using the data after the deletion as learning data and stores the learning result information in a learning result DB (32). The signal intensity of the predetermined m / z value is likewise deleted also with regard to the mass spectrum data relating to the target sample to be determined, and then a determination processing unit (24) determines the classification into the class corresponding to the presence or absence of cancer on the basis of the learning result information stored in the learning result DB (32).

Description

technical field [0001] The present invention relates to a data analysis device and analysis method for analyzing data obtained by a mass spectrometer, and more specifically, to diagnosis or inspection of specific diseases, authenticity determination of origin of agricultural products and fishery products, and counterfeit medicines. A mass spectrometry data analysis device and analysis method useful for inspection and determination of samples in various fields such as identification of counterfeit money and the like. Background technique [0002] With the rapid development of mass spectrometry technology in recent years, attempts have been made to perform mass spectrometry analysis on biological samples (blood, urine, saliva, or a part of biological tissue, etc.) Data analysis of the data to diagnose specific diseases such as cancer. For example, Patent Document 1 of the present applicant et al. describes the following content: dPLRM (dual Penalized Logistic Regression Machi...

AI Technical Summary

Problems solved by technology

Therefore, there is a problem that the determination of cancer based on such learning results and the examination of tumor markers targeting these marker molecules cannot be said to be independent from each other, and it is difficult to ensure the accuracy and reliability of diagnosis combining them.

[0007] In addition, due to various factors such as the following, there may be fluctuations far exceeding the range of data fluctuations caused by individual differences in subjects during learning and judgment, or it may be difficult to obtain data for learning data and data for judgment. The conditions are the same, which becomes a reason for reducing the accuracy of the judgment

[0008] (1) Drugs such as various anticancer drugs are usually administered to cancer patients during the course of cancer treatment, but changes in biological tissue caused by such treatment appearing in the learning data may be erroneously learned as originating from cancer organizational changes

[0009] (2) Contrary to (1), in the case of judging data obtained from a patient who was administered a new anticancer drug not included in the learning data based on the learning result, when the occurrence of cancer and non-cancer drugs on the mass spectrum When a peak derived from a new anticancer drug occasionally exists near the mass-to-charge ratio of the resulting difference, it may be misidentified as a cancerous / noncancerous tissue change

Similarly, when there is a peak derived from a foreign matter or a foreign matter not included in the learning data in the data for cancer / non-cancer determination, it may be misjudged as a tissue change derived from cancer / non-cancer

In addition, internal standard substances may be added to the sample to be judged, but peaks derived from internal standard substances may be misjudged as cancerous / non-cancerous tissue changes

[0010] (3) Biological samples from cancer patients and healthy individuals are usually collected and measured by medical institutions such as hospitals, but it is difficult to collect / measure samples with the same procedures and quality in each of these institutions

However, recollecting and measuring multiple biological samples that meet the target is a very laborious and time-consuming operation, and in most cases it is virtually impossible

[0012] In addition, although the factors caused by the sample being a sample derived from a living body are particularly large for the above-mentioned problems, the same or similar problems may also occur in fields other than disease diagnosis and the like.

Images