TY  - JOUR
N1  - Open Access
ID  - open1860
UR  - http://www.nature.com/articles/srep22843
A1  - Chaudhary, Kumardeep
A1  - Kumar, Ritesh
A1  - Singh, Sandeep
A1  - Tuknait, Abhishek
A1  - Gautam, Ankur
A1  - Mathur, Deepika
A1  - Anand, Priya
A1  - Varshney, Grish C
A1  - Raghava, G.P.S.
Y1  - 2016///
N2  - Numerous therapeutic peptides do not enter the clinical trials just because of their high hemolytic activity. Recently, we developed a database, Hemolytik, for maintaining experimentally validated hemolytic and non-hemolytic peptides. The present study describes a web server and mobile app developed for predicting, and screening of peptides having hemolytic potency. Firstly, we generated a dataset HemoPI-1 that contains 552 hemolytic peptides extracted from Hemolytik database and 552 random non-hemolytic peptides (from Swiss-Prot). The sequence analysis of these peptides revealed that certain residues (e.g., L, K, F, W) and motifs (e.g., "FKK", "LKL", "KKLL", "KWK", "VLK", "CYCR", "CRR", "RFC", "RRR", "LKKL") are more abundant in hemolytic peptides. Therefore, we developed models for discriminating hemolytic and non-hemolytic peptides using various machine learning techniques and achieved more than 95% accuracy. We also developed models for discriminating peptides having high and low hemolytic potential on different datasets called HemoPI-2 and HemoPI-3. In order to serve the scientific community, we developed a web server, mobile app and JAVA-based standalone software (http://crdd.osdd.net/raghava/hemopi/).
PB  - NPG
JF  - Scientific reports
VL  - 6
KW  - Computational models;Machine learning;Software
SN  - 2045-2322
TI  - A Web Server and Mobile App for Computing Hemolytic Potency of Peptides.
ER  -