COMSATS Institute of Information Technology, Islamabad
Sunday, May 20, 2012
Bio Inspired Advance IT Group
Introduction
The umbrella activities in the group signifies the move from the early focus on the isolation and identification of specific genes behavior among heterogeneous disease in IT respects, including its cellularity, different genetic alterations and diverse clinical behaviors. Capitalizing on High Performance Computing (HPC) machines and the wealth of in-house developed algorithms, DNA missing sequences prediction and Microarray data repairing, to name a few, a compute-intensive scientific research can be followed for high throughput technologies on Dry-Wet project basis that will help researchers shed light on future scientific questions.
Members
Alumni
  • Ms. Sobiah Rauf
  • Ms. Sidra Batool
  • Ms. Nosheen Riaz
  • Mr. Muhammad Asad
  • Ms. Mehwish Naqvi
  • Ms. Mehwish Huma
  • Ms. Syed Saud Naqvi
  • Ms. Hina Khalid
  • Ms. Asmara Khan
Courses Offered
 
(MATLAB for  Bio-informatics and Bio-Sciences) (Biomedical Imaging) (Modeling and Simulations)
Click here
Available projects for MS, BS students
• A computational tool for integrated visualization of Protein interaction networks
Hasan Bilal Mirza, Ayesha Hameed, Naeem Zafar Azeemi
Progress in the reliability and throughput of protein physical interaction detection techniques (both experimental-and computational-wise is gradually leading to the availability of more comprehensive, higher confidence protein interaction data. These maps can serve as invaluable tools for biological research, in particular for more integrated system-level studies of biological processes and mechanisms. We aim to develop a computational application for exposing the architecture of protein interaction networks. It facilitates the system-level analysis of mRNA expression data in the context of the underlying protein interaction network. We want to preliminary analysis of a human protein interaction network and comparison with other species.
• Transcription networks modeling with amorphous computations
Ayesha Hameed, Hasan Bilal Mirza, Naeem Zafar Azeemi
Chemists and biologists are in the process of creating many implementations of chemical computation and control logic. The power of electronic computation is due in part to the development of modular gate structures that can be coupled to carry out sophisticated logical operations and whose performance can be readily modeled. We aim to develop the equivalences between electronic and biochemical operations at gate level. In order to help cross between these disciplines, we want to develop an analogy between complementary metal oxide semiconductor (CMOS) and transcriptional logic gates. Our objective is to implement these computations, by using various design biological system level gates and then characterize these gates in a binary latch similar to that already .
Current Research Activities
• Gene Disease Relationship
Sobiah Rauf, Sidra Batool, Nosheen Riaz, Ayesha Hameed, Naeem Zafar Azeemi.
A disease is an abnormal condition of an organism that impairs bodily functions, associated with specific symptoms and signs. It may be caused by external factors, such as infectious disease, or it may be caused by internal dysfunctions, such as changes in normal gene function.This project is based on gene disease relationship in which we will categorize various diseases for e.g intrinsic, extrinsic, infectious, non-infectious, diseases of unknown Origin, incurable diseases and different disorders, to name a few. We will study different organisms in which cancer is common according to genes e.g. genes causing cancer in humans, zebra fish and mouse. We will use different association models of bioinformatics for cancer study i.e. SNPStats, Electronic PCR, Entrez Gene, Model Maker, ORF Finder, SAGEmap, Spidey, Splign, COGs, Clusters of Orthologous Groups. Our survey about diseases will include /literature review/ i.e. different journals including Oncology Journal, Cancer Journal for Clinicians, Bioinformatics, Bioscience horizons, Briefings in Bioinformatics, Human Molecular genetics, DNA research etc and /population based study/. We will also find different factors and sub factors involved in gene to diseases pathway and their relationship (to each other and with pathway). Finally we will develop association models using UML (Unified modeling language).
• Building UML model for vitamin-D deficiency associated gene-disease system
Najia Nawaz, mehwish Naqvi, Ayesha Hameed, Naeem Zafar Azeemi.
Vitamin D is a group of fat-soluble prohormones, obtained from sun exposure, food, and supplements. It is biologically inert and must undergo two hydroxylation reactions to be activated in the body. Active form of vitamin D plays major role is to increase the flow of calcium into the bloodstream, by promoting absorption of calcium and phosphorus from food in the intestines. It mediates its biological effect by binding to Vitamin D Receptor (VDR) located in nuclei of target cells and allows it to act as transcription factor for expression of various genes. Common diseases caused by vitamin D deficiency may include cardiovascular diseases like blood pressure, heart attack, etc. infectious diseases like influenza and respiratory diseases like asthma, tuberculosis. In this work we shall classify vitamin D deficient diseases with respect to their causes e.g. cause may be genetic, environmental and nutritional. Then population based study of relation between vitamin D and diseases will be performed. Case study will be of genetic behavior of certain diseases. Finally a UML based model shall be developed for vitamin D deficient gene disease association system.
• Detection and Classification of Masses in Mammogram
Mehwish Huma, Dr Naeem Zafar Azeemi.
Breast cancer is the second most common cause of morbidity and mortality in women (American Cancer Society, 2008). Mammography screens the breast to differentiate with four levels of the intensities such as background, fat tissue, parenchyma and calcifications with increasing intensities. In radiologist viewpoint, visual interpretation of mammograms is considered to be a very demanding job. Their judgment essentially depends on the training, experience and subjective criteria. Thus, it is important to develop Computer Aided Diagnosis (CAD) that can distinguish benign and malignant lesions effectively to improve the rate of detection accuracy of masses. In this work, we focus on low contrast mammograms that make it more difficult to detect masses than micro-calcifications, since the low-level features present normally found to be obscured or similar to normal breast parenchyma. In addition, the masses are quite thin and often present in the dense areas of the breast tissue. It has smoother boundaries than micro-calcification and has shapes like circumscribed, speculated, lobulated or ill-defined. The circumscribed ones usually have distinct boundaries of 2-30 mm in diameters and high-density radiopaque. Among these, the speculated ones have rough, star-shaped boundaries and the lobulated ones have irregular shapes. The masses present a great challenge to researcher to be identified as benign or malignant for a CAD system to improve the biopsy yield ratio. Further, the masses shall be classified as malignant and benign based on certain properties of the respective features known as descriptors. While the masses with radiopaque and more irregular shapes shall also be explored for false positive or negative signatures.
Completed Research Activities
• Reference Free Framework for Bio-Inspired Real-Time Motion Detector
Syed Saud Naqvi, Naeem Zafar Azeemi.
This work proposes a framework for identification of moving objects by incorporating primary response channels of the retina in patients suffering from degenerative defects. The goal is to significantly revive the primary visual sensations such as directional movement detection and differentiation of static and moving objects. A Biological Neural Network (BNN) is proposed that can identify the direction of moving object based on the concepts of neural coding and Mean Firing Rate (MFR). Our model is based on Spiking Neuron Models (SNM) that is close to leaky integrating and fire principle and the Hodgkin-Huxley principle. Proposed neural network architecture performs directional movement detection depending on spiking behavior of neuron groups. It is capable of identifying movements in any direction depending on the delay profile of identical spiking behavior observed between two distant neuron colonies. Our results show that, proposed network model resembles the actual motion processing visual path way of the human retina.
• Missing Value Prediction in Microarray Asymmetric datasets
Qurrat-ul-Ain, Hina Khalid, Ayesha Hameed, Naeem Zafar Azeemi.
DNA microarray technology has become a powerful tool in the arsenal of the molecular biologist. Microarray experiments are performed to gather gene expression datasets. The reliability of datasets is hampered by many debilitating factors, such as instrumentation error, missing values, to name a few. While the fabrication of microarrays is a very labor-intensive process, production of quality microarrays individually tailored on a project-by-project basis will help researchers shed light on future scientific questions. In the same vein, regardless of the printing arrayer used, the calibration of the robot is absolutely critical. The tolerances needed are very tight, and every micron out of 'true' can cause problems. In this work we use non-linear estimation technique to predict missing values in the asymmetric datasets. Reliability is observed by comparing Normalized Root Mean Square Error of Prediction (NRMSEP) on number of Principal Component (PC) for data estimation of DNA microarray. Non-Linear PCA (NLPCA) technique is applied to check the behavior of non-linearly structured datasets. Results are exposed for Bayesian PCA (BPCA), Non-Linear PCA (NLPCA), Local Least Square Impute (LLSImpute), Single Value Decomposition (SVD) Impute, Non-Linear Iterative Partial Least Squares and Probabilistic PCA (PPCA). We achieve best estimation results in BPCA and PPCA for large gene expression datasets. The robustness of proposed methodology is verified for missing values prediction in datasets available at public websites, such as HeLa (marker genes of human cell cycle), metabolite data (plant Arabidopsis thaliana).
• Impact of Factor Selection on the First Overtone Transmission Spectra in Optical Glucometer
Syed Rameez Naqvi, Naeem Zafar Azeemi.
Glucose monitoring devices use Multivariate Calibration (MC) methods to estimate glucose concentration in blood. The accuracy of methods depends on spectral data obtained from tongue-to-spectrometer interface. In this work we examine four widely used MC methods, they are: Classical Least Square (CLS), Inverse Least Square (ILS), Principal Component (PC) and Partial Least Square (PLS). We discuss the impact of factor selection on the prediction of response for first overtone transmission spectra collected across human tongues. Results are exposed for various Signal-to-Noise Ratio (SNR) and different factor values. We explicitly tackle the issue of low SNR in tongue-to-spectrometer interface. We show that CLS outperforms factor based regression techniques where SNR is as low as 30 dB. Our results are useful in calibration models that are used to predict vivo glycemia from human tongue spectra.
• Improving Accuracy of Non-Invasive Glucose Monitoring Through Non-local Data Denoising
Syed Rameez Naqvi, Naeem Zafar Azeemi, Ayesha Hameed.
Correlation and clinical interpretation, in respect to the true glucose value of patient is imperative for optimum therapy and disease management. Accuracy of optical glucometer is hampered by many debilitating factors such as concentration range, sampling environment, tongue-to spectrometer interface, changes in wavelength, polarization or intensity of light, to name a few. Regression techniques are used in such devices to build patient glucose model. This work is an extension to our previous work regarding ltivariate calibration for glucose level prediction in non-invasive human tongue spectra. Here, we present our results for noise reduction and data conditioning during glucose spectrum isolation phase. We embed our ‘Indicator Function (IF)’ scheme into two popular techniques known as Outlier Sample Removal (OSR) and Descriptor Selection (DS). Methodology is tested on dataset 'OCATNE20' obtained from a public domain website and results are compared at both OSR and DS for a wide range of blood serum samples. Our results show that outlier samples identification and removal in early stage significantly increase the prediction of unknown samples typically in the range of 7.95% to 9.84%.
• Ultra Wide Band Radar Based Tamper-Resistant Clinical Asset Tracking System (ATS)
Asmara Khan, Naeem Zafar Azeemi.
From a patient safety perspective, time-consuming searches for life saving equipments or drugs have been the subject of a hazard report or recall. In addition to yielding significant time saving Asset Tracking Systems (ATS) ensure all assets tracking. In this work, an Ultra Wide Band (UWB) tracking system is developed, that automatically captures the movement of in-house patients and clinical stuff. The whole system comprises of Position Assessment Block (PAB), Position Management Block (PAB) and Application Processing Unit (APU), connected through UWB and Wireless LAN network. System is tested in a typical hospital ward environment for diversified UWB receiver tracking methodologies, such as Point Mode Tracking (PMT), Linear Scan Mode Tracking (LSMT), Bat Mode Tracking (MBT) and Panoramic Tracking (PMT). Characterization of suitable tracking schemes is also discussed. Robustness of system is verified for across various position detection algorithms for e.g., Angle of Arrival (AoA), Received Signal Strength (RSS) and Time of Arrival (TOA). In addition to helping staff locate equipment, our ATS is able to facilitate use and distribution of inventory by collecting and facilitating the analysis of data on equipment location over time and hence prone to be tampered.
• On-Site Ultra Wide Band Construction Material Tracking System (UWB-CMTS)
Asmara Khan, Naeem Zafar Azeemi
On-site building material and construction tools management offers a big challenge to traditional asset tracking systems. Construction managers determine the best way to track building material, construction tools and equipment to rationally layout the most cost-effective plan and schedule for completing the project. In contrast to conventional asset tracking schemes such as Radio Frequency IDs (RFID), Infra Red IDs (IRID), Tagged IDs (TID) to name a few, we use Ultra Wide Band (UWB) technology in our UWB Construction Material Tracking System (UWB-CMTS), that is relatively new and gaining researchers interest in imaging as well as in tracking. We transmit short impulses ranging from sub-nanoseconds to a few nanosecond are transmitted that have high material-penetration capability, low electromagnetic interference as well as lower specific absorption rates and good immunity against multipath interference. We model a typical building system into many segments such as structural framework, floors, and walls, including fire-protection, electrical, plumbing, air-conditioning, and heating. This methodology made our scheme well suited for geometry or contour based construction asset location estimation. We explore variability of tracking schemes for number of Ultra Wide Band (UWB) based radar tracking systems. Results are exposed for number of construction scenarios; UWB antennas were placed taking into accounts various UWB tag placements and reader locations. Robustness of system is verified across various position detection algorithms for e.g., Angle of Arrival (AoA), Received Signal Strength (RSS) and Time of Arrival (TOA). The system is expected to be useful in a large-scale construction machinery movement especially for construction projects, such as an office building or industrial complex.
Log In Information
User Name:
Password: