Proteomics

Materials and methods

Protein cluster analysis

First, the quantitative information of the target protein set is normalized (normalized to the interval (-1,1)). Then, The Complexheatmap R package (R Version 3.4) was used for both sample and protein expression dimensions Classification (distance algorithm: Euclidean, linkage method: Average linkage), and generate hierarchical clustering heat maps.

Subcellular localization analysis

The CELLO (http://cello.life.nctu.edu.tw/), the method of location prediction of cells, the method is adopted multi-class SVM (support vector Machine) machine learning methods are known for subcellular localization information in public databases Protein sequence data modeling is used to predict the subcellular localization information of the protein to be retrieved.

Protein domain analysis

Protein domain analysis uses the Pfam database, which is a collection of a series of protein families, each of which Each protein family is represented in the form of multiple sequence alignment and hidden Markov model. This database contains domain information For specific analysis, use the interproscan package (version number: InterproScan-5.25-64.0) to integrate By running a scanning algorithm from the InterPro database for functional characterization of sequences, the target protein sequence is obtained in Pfam Domain annotation information in the database.

GO function annotation

Blast2GO (version number: BLASTP 2.8.0+) was used to annotate the target protein set It can be summarized as sequence alignment (BLAST), GO entry extraction (Mapping), GO Annotation (Annotation) and InterProScan adds four steps to Annotation Augmentation.

KEGG path annotation

KOBAS(version: KOBAS 3.0) software was used to annotate the KEGG pathway of the target protein set.