Amyloid

a platform for protein sequence analysis based on machine learning approaches

Input protein sequences below:
The format of this site should be fasta format.
The first line is any text that starts with ">" and the analysis only works for fasta header with "|" character and labels.
Starting from the second line is the sequence itself, allowing only the use of established amino acid encoding symbols.


You can paste the sequence into the following table(or use the example ) or Upload data file




Welcome to RFAmy server

This paper uses 188-D feature representation with the best performance according to the experimental results. The method entails the following main steps. First, the original protein sequence was generated from the Uniprot and AmyPro datasets and then subjected to a de-redundant operation to generate the final protein sequence data called Amy. The second step is feature extraction of protein sequences. The third step is to use RF to classify protein sequences. Figure 1 shows the experimentalprocess in this paper.
the experimental process in this paper

Figure 1. Overview of the paper framework for a Amyloid classifier.

Datasets

This article uses a self-built data set named Amy. The source databases for protein sequences are the Universal Protein (UniProt, http://www.uniprot.org/)and Amyloid (AmyPro, http://www.amypro.net/)databases. It consists of 165 amyloid proteins(positive samples) and 382 non-amyloid(negative samples).

1. The dataset Amy with FASTA format can be download (Here).

2. The feature set corresponding to the dataset Amy can be downloaded(Here).

Help

The site is for the identification of protein sequences to determine whether it is amyloid protein. The format of this site should be fasta format.The first line is any text that starts with ">" and the analysis only works for fasta header with "|" character and labels. Starting from the second line is the sequence itself, allowing only the use of established amino acid encoding symbols.
e.g
>2MA1A|1
HDAPLFEALRAWRLQKAKELSLPPYTIFHDATLKTIAELRPGSHATLGTVSGVGGRKLAAYGDEVLQVVRDSSGG

Contact

Dr. Quan Zou
Tianjin University, School of Computer Science and Technology, China
Email: zouquan@tju.edu.cn
Personal website:http://lab.malab.cn/~zq/