A webserver for classifying the molecular functions of Rab GTPase using Deep Learning
Introduction
Rab protein is a member of monomeric G-protein, which regulates numerous essential process in the cell. Rab proteins
were usually bound with GTP binding sites to perform various functions in biology molecules. According to the functional
role in membrane trafficking and cell physiology, Rab proteins could be classified into four classes, they are Rab
GDP-dissociation inhibitor (GDI) activity, Rab geranylgeranyltransferase activity (GGT), Rab GTPase binding, Rab guanyl-nucleotide
exchange factor activity. Fig. 1 shows the process of Rab cycle including all the molecular functions. Rab GGT plays
the role as an enzyme in Rab system to transfer the geranylgeranyl and deliver the Rab to its target membrane. GDI
activity prevents the association of GTP binding in Rab proteins system, so that keeps the Rab proteins in inactive
state. Rab GEF activity stimulates the exchange of GDP to GTP binding sites to activate the Rab. Under normal condition,
the priority of GTP is higher than GDP, then Rab GEF activity is activated to replace GDP by GTP. A GTPase binding
aims to interact Rab with the binding sites and molecules and thereby convert the Rab back to its inactive state. The
Rab then interacts with GTPase activating protein (GAP) and removed from the membrane via GDI.
Numerous types of Rab proteins now identified in human and several studies demonstrated that a functional loss of Rab
proteins at each of the membrane trafficking steps had been implicated in many diseases i.e., choroideremia, intellectual
disability, cancer, Parkinson’s disease. For instance, the modulation of Rab GGT molecular function has been related
to Hermansky-Pudlak syndrome. The Rab GGT also used to develop some drug targets related to bone diseases, such as
osteoporosis. Mutations in the Rab3GAP, which is a human class of the GAP process lead to X-linked nonspecific mental
retardation, Warburg Micro and Martsolf syndromes, diseases characterized by developmental abnormalities of the eye,
nervous system, and genitalia. A Rab GEF has also been implicated in human disease. Thus classification of Rab protein
is a crucial topic and there is a requirement to develop some computational techniques to identify them.
Method
We approached a precise model using 2D CNN and PSSM profiles to classify the Rabs molecular functions in membrane trafficking.
The flowchart of the study included four subprocesses: data collection, feature set generation, CNN generation and
model evaluation. We describe the details of the proposed approach as follows.
Dataset
All the dataset using in this web server are retrieved from UniProt and GeneOntoly. The detail of the dataset lists in
the below table.
Original
BLAST 30%
Cross-validation
Independent
GDP-dissociation inhibitor
2060
169
141
28
geranylgeranyltransferase
1882
185
155
30
GTPase binding
10014
2694
2157
431
guanyl-nucleotide exchange factor
3136
630
525
105
If you would like to build a model and evaluate our model, we provide the dataset as the below link.
In order to avoid the errors, please submit the sequence in fasta format (we also give you the fasta file examples).
The user can choose two options to submit, including paste the sequence into text area and upload sequence file.
The user can submit one single fasta file or multiple fasta file. In the result page, the probabilities will be
shown to help people choosing which protein belongs to the corresponding complex of electron transport chain.
Department of Computer Science and Engineering
Yuan Ze University
135 Yuan-Tung Road, Chung-Li, Taiwan 32003, R.O.C.
Nguyen-Quoc-Khanh Le
Research Scholar
Department of Computer Science and Engineering
Yuan Ze University
135 Yuan-Tung Road, Chung-Li, Taiwan 32003, R.O.C.
Quang-Thai Ho
Research Scholar
Department of Computer Science and Engineering
Yuan Ze University
135 Yuan-Tung Road, Chung-Li, Taiwan 32003, R.O.C.
Contact us
Yuan Ze University Department of Computer Science and Engineering
Graduate Program in Biomedical Informatics
Bioinformatics Laboratory (R1607B)
Address: No. 135, Yuandong Road, Chungli City, Taoyuan County, Taiwan R.O.C .32003
Tel: (03) 463-8800
If you have any problem or suggest any idea for our website, feel free to contact us via email:
[email protected]