Dr Muhammad Saqib Sohail

PhD Theme/Topic: A generalized Human Leukocyte Antigen (HLA-peptide binding prediction framework

Supervisor: Dr Muhammad Saqib Sohail/ Snr. Associate Professor
Email: saqibsohail.bulc@bahria.edu.pk
Supervisor’s google scholar profile: https://scholar.google.com/citations?user=XRuF7HAAAAAJ&hl=e

Campus/School/Dept: BULC /-/CS

Topic Brief Description:

Predicting Human Leukocyte Antigen (HLA)-peptide binding is central to the development of vaccines, cancer immunotherapies, and has applications in personalized medicine. Current HLA-peptide binding prediction methods frequently yield inconsistent peptide rankings and exhibit errors that reflect differing underlying feature representations. This project aims to address these limitations by developing a novel generalized HLA-peptide binding prediction framework that leverages the strengths of current approaches. By employing ensemble learning to synthesize insights from multiple models and utilizing transfer learning, the proposed framework seeks to enhance prediction performance.

Research Objectives:
To create a generalized framework for HLA-peptide binding prediction that enhances performance by developing a state-of-the-art method that also incorporates insights from existing methods.

Deliverable:

1. A trained machine learning model for HLA-peptide binding predictions.
2. A webserver with a user-friendly graphical user interface that the public can use to input a set of peptides and HLAs and predict their binding affinity using the developed model.

Research Questions: 

1. How can the available HLA-peptide binding data be used to develop a generalized machine learning method for HLA-peptide binding predictions?
2. What factors affect the types of miss-classifications made by existing HLA-peptide binding prediction methods and how can these by leveraged to improve model predictions?

Why you should apply:

1. You get to develop state-of-the-art machine learning models for predicting HLA-peptide binding predictions.
2. You get to work with a highly motivated international team of researchers.
3. Excellent opportunity to develop your technical and soft skills.
4. You get to do research that has the potential to have a big impact on vaccine design, cancer immunotherapies, and the emerging field of personalized medicine 

Candidate’s Eligibility Profile:

1. The applicant must have an MS/MPhil/Equivalent degree in CS/CE/EE with CGPA > 3.0. An ideal candidate would have a strong background in Machine learning and artificial intelligence. Additional understanding of detection and estimation/signal processing would be a plus.
2. Good programming skills are a must (experience in any of Python, R, or C/C++). Experience in developing/training machine learning/artificial intelligence models is desirable.
3. Candidates with prior publications are strongly encouraged to apply.
4. Excellent communication skills.
5. Proficiency in spoken and written English is essential.
6. We value independence while promoting teamwork and collaboration with colleagues.

===============================================================================================================

PhD Theme/Topic:  An explainable AI-based system for Automated X-Ray Report Generation

Supervisor: Dr Muhammad Saqib Sohail/ Snr. Associate Professor
Co-Supervisor: Dr Khawaja Qasim Maqbool/ Associate Professor
Email: saqibsohail.bulc@bahria.edu.pk

Supervisor’s google scholar profile: https://scholar.google.com/citations?user=XRuF7HAAAAAJ&hl=en

Campus/School/Dept: BULC /-/CS

Topic Brief Description:

Government hospitals in Pakistan face significant challenges due to a shortage of doctors and overwhelming patient loads, particularly in radiology departments. This creates substantial challenges in patient care. Long delays in the evaluation of X-ray images and the generation of diagnostic reports often hinder timely treatment. This project proposes an explainable AI-based system for automated X-ray report generation. The system will leverage state-of-the-art artificial intelligence methodologies to analyze X-ray images and produce accurate diagnostic reports. Additionally, we will identify the features learned by the model and their impact on accurate X-ray classification and reporting. By automating the reporting process, the framework aims to alleviate the burden on medical professionals, significantly reduce patient wait times, and enhance operational efficiency within hospitals. With its scalable and adaptable design, the proposed solution will be tailored to meet the pressing demands of the radiology departments of Pakistan’s overburdened public healthcare system.

Research Objectives:

To develop a scalable and adaptive AI-based framework for automated X-ray report generation to enhance diagnostic efficiency and alleviate the workload in Pakistan's public healthcare system.

Deliverable: 

A fully functional AI-powered software prototype capable of analyzing X-ray images and generating diagnostic reports in real-time.

Research Questions: 

1. How can historical X-ray data be used to train an explainable AI-based system for automated report generation?
2. How does the choice of model architecture affect system performance?

Why you should apply:

1. Opportunity to do research work on cutting-edge technology. You will be designing explainable models for classifying X-Ray scans.
2. Community-focused project solving a real-world problem faced by ordinary Pakistanis.
3. Opportunity to work with a highly motivated team with international experience.
4. Excellent opportunity to develop your technical and soft skills.

Candidate’s Eligibility Profile:

1. The applicant must have an MS/MPhil/Equivalent degree in CS/CE/EE with CGPA > 3.0. An ideal candidate would have a strong background in Machine learning and artificial intelligence. Additional understanding of detection and estimation/signal processing would be a plus.
2. Good programming skills are a must (experience in any of Python, R, or C/C++). Experience in developing/training machine learning/artificial intelligence models is desirable.
3. Candidates with prior publications are strongly encouraged to apply.
4. Excellent communication skills.
5. Proficiency in spoken and written English is essential.
6. We value independence while promoting teamwork and collaboration with colleagues.