Deep learning under mass-to-charge ratio pre-retrieval to realize electron ionization mass spectrometry library retrieval.

RATIONALE: Gas chromatography-mass spectrometry (GC-MS) is an analytical technique widely used in materials science, biomedicine, and other fields. The target compound in the experiment is identified by searching for its mass spectrum in a large mass spectrum database using some algorithms. This work introduces the use of deep learning ranking for the identification of small molecules using low-resolution electron ionization MS. Because different spectra are often very similar, the algorithm produces wrong search results, and the search accuracy needs improvement. Due to the library's large amount of data, the algorithm sometimes requires a large amount of calculation and is very time consuming. METHODS: Given these two problems, this work aims to develop a model for ranking based on mass-to-charge ratio (m/z) pre-retrieval method combined with deep learning to improve search accuracy and reduce the algorithm's computational time. The master spectral library maintained by the National Institute of Standards and Technology is used as the reference library for all the experiments, and the replicate library is used as the query library to evaluate the method's performance. RESULTS: Compared with non-machine learning algorithms, the combination of m/z matching pre-retrieval and deep learning significantly improves library retrieval accuracy by about 4%. Moreover, compared with the deep learning sorting algorithm that does not use the pre-retrieval process, it improves the accuracy of spectral library retrieval by about 0.1% and reduces the computational time of the algorithm by more than 2 h. CONCLUSIONS: This method identifies compounds more efficiently and accurately than non-machine learning and deep learning algorithms without a pre-retrieval process.

[Urban land use change detection based on high accuracy classification of multispectral remote sensing imagery].

In the present paper, the urban land change in Jiading district of Shanghai was studied on the basis of high accuracy classification for 4 epochs of multispectral remotely sensed imageries. A further improved genetic-algorithm optimized back propagation neural network approach was first employed in our study to obtain sorts of land cover types from the remotely sensed imageries. The urban land and non-urban land types were thus extracted based on the classification result. According to the 16 corresponding relationships between the pixel values in the four urban land imageries and the ones in the generated urban land change imagery, the amount of each type pixel in the generated imagery was calculated according to the four plates, and the situation of urban land change was analyzed and investigated for the study area in three year intervals. The urban development in the study area was also preliminarily revealed.