GSoC
GSoC copied to clipboard
cBioPortal
Patient Similarity
Background: Given the genomic and clinical data of a patient, finding patients with similar genomic and clinical profiles is potentially very useful for disease diagnosis or rediagnosis diseases and treatment decisions based on previous similar patients.
Goal: The overall goal is to develop a new tab Similar Patients in Patient View (e.g. a TCGA endometrial patient) listing similar patients within the study. It is important that loading the similar patients is done quickly.
Approach:
- Define a metric of similarity between two patients. A simple metric is the number of common genomic alterations. Another one is Jaccard Index used in this paper.
- A slightly more advanced calculation would be to use a weighted calculation. Weighted correlation R code and example input data. NOTE: Weights would be provided by mentors
- Develop an API for getting similar patients for one patient. The similarity matrix could be precalculated and saved into database for a study.
- This sites may provide applicants some with inspiration http://projects.sanderlab.org/tumorcomparer/ and https://depmap.org/portal/celligner/
- The most simple interface modification that could be imagined is to add a new tab to Patient pages, example patient page, that says "Similar Patient". The new tab would have a table with two columns: PatientID and OverallSimilarityValue. The PatientID would link to another patient in cBioPortal.
Need skills: Java, Javascript, ReactJS
Possible mentors: @cannin
Hi ! I am a final year Undergrad student at IIT Kharagpur, India.
From a first read of the above mentioned work by Rileen Sinha et al. I have briefly understood the potential of this project. The weighted similarity based approach (TumorComparer) which has been arrived to, by statistical analysis (kNN) can be used in comparisons of genomic profiles of patients. The results of which can be promising in providing better and a personalized cancer therapy.
I am really interested to work towards adding this feature of Patient Similarity to the Patient View of cBioPortal. I have experience working with Java, have a basic knowledge of statistical analysis and numerical computing and am inclined to learn any of the technologies that might be required in making this feature as effective as possible.
It would be great if @inodb could connect me with the possible mentors so that I can discuss the right way to proceed.
@cAbhi15 thanks for your interest. To get started, please see the how to apply section. If you have specific questions, you can comment here or ask us on Slack
@jjgao Hello Mentor. This is Vinay Sagar Sharma, currently doing my B.tech in Computer Science Engineering from Shri Mata Vaishno Devi University, Katra, India. I am interested to participate in GSoC 2019. I am interested to work with cBioPortal as I found it fascinating and the tech stack was just what I was looking for to work with. I hope it would be a happy journey and fun coding this summer with cBioPortal if selected. My Tech Stack includes HTML,CSS,Javascript(both ES5 and ES6). I use React as a front-end framework and I have also worked on Redux. I am good with Python . I found cBioPortal really fascinating and would love to be part of this community. Looking for some guidance to kick off. :smile: Thank you
Hi mentor @jjgao. I am Senuri De Silva, a final year student from University of Moratuwa, Sri Lanka. I am very interested in this project and I have gone through the tutorials of cBio Portal and currently in the process of familiarizing myself with the code base. I have worked with ReactJs, Javascript, and Java in my internship and other projects. I personally love to work in biomedical research and I suppose this is a very good opportunity to follow my passion. I hope my journey with cBio will be an exciting one. I have submitted my draft proposal to the GSoC portal and your feedback is highly appreciated. Thank you!
@jjgao Hi! I want to work on this project. Can you tell me one thing? Is this project coming in GSoC,2020?
@sunil-17112 : @Luke-Sikina has been working on developing a prototype. I'm not sure if this idea is suitable for GSoC2020. What do you think @Luke-Sikina ? It might be a bit too early in the prototype process.
@inodb I think I have to learn more about what constitutes a good GSoC project. The spec in this document is very open ended, where as our current vision is far more specific. Are students interested in issues where the creative process is more limited? If I were a student, yes, I think this would be a good project. I think the narrower scope and more coherent expectations makes it a more gratifying project and more likely to succeed. It needs a lot of rewriting to reflect the current status of similar patients, but I can work on that.
Hi, could you please assign this issue to me? I would love to work on it and contribute to the project
@lalitkumawat1m Thanks for the interest. Make sure to look at the "How to apply" (https://github.com/cBioPortal/GSoC) information for your application.
