Sequence analysis method, sequence analyzing device, polymerization condition proposal device, and automatic synthesizing device

A polymer sequence analyzing device for inferring a polyad content of a polymer obtained by polymerizing monomers selected from a monomer set containing two or more types of monomers, which polyad is formed by aligning a plurality of units derived from the monomers, the device comprising:

• a mass spectrometer that sequentially ionizes gas components generated by heating a sample comprising a reference sample and an inference target sample, which are polymers formed from the monomers, and observes mass spectra continuously, and
• an information processing device that processes the observed mass spectra, wherein
• the information processing device comprises:
• a data matrix generating part that obtains a data matrix comprising two-dimensional mass spectra having m/z with respect to heating temperature,
• a variant number determining part that determines number K of variants of the polyads depending on number of types of monomers contained in the monomer set and number of the units forming the polyads,
• a first NMF processing part that performs a first NMF process by which the data matrix is subjected to non-negative matrix factorization to be factorized into a product of a matrix representing a normalized basis spectrum and an intensity distribution matrix,
• a second NMF processing part that performs a second NMF process by which the intensity distribution matrix of each of the samples is subjected to non-negative matrix factorization to be factorized into a product of a matrix representing a mass proportion of a model polymer formed only from the polyad in the sample, and a matrix representing a feature vector of the model polymer and obtains the feature vector of the model polymer,
• a vector projection part that defines the feature vector of the model polymer as an end member and determines a K−1 dimensional simplex comprising all of the feature vectors of the samples, and
• a composition inference part that defines a distance between the K end members and a feature vector of the inference target sample by Riemannian metric distance in consideration of non-orthogonality of a basis spectrum of first NMF process, and infers each polyad content ratio of the inference target sample by a ratio of the distance.