Week 6 Updates: ClinPrior Candidates and Cognition Lecture

Caitlin E -

Hi Everyone! This week introduced me to a plethora of new information. From finding candidate gene variants with ClinPrior to listening to a lecture on cognition and aging, what I learned this week has brought me closer to answering my original research question. 

AI Tools 

At this point, the Schrauwen laboratory sees little hope for AI MARRVEL. Besides prioritizing a variant’s allelic frequency over its effects, MARRVEL generally scans intronic variants more than exonic variants. Since intronic variants are non-coding, they are of minimal interest when matching gene variants to specific phenotypes. After MARRVEL produced multiple inaccurate readings and rankings, our efforts have shifted primarily to ClinPrior and its applications. 

So far, ClinPrior has produced some interesting leads on the potential genetic diagnoses for neurological diseases. The UGDH gene, for instance, has been linked to epilepsy, among other developmental issues. The results from ClinPrior were insightful; however, it was necessary to look to outside sources to confirm the phenotypes associated with certain genes and gene variants. OMIM, or the Online Mendelian Inheritance in Man, was an extremely useful database in learning about genes, gene variants, and specific phenotypes related to either. In the next few weeks, I hope that my laptop does not RESTART UNEXPECTEDLY and take an ENTIRE HOUR to update right in the middle of my work session as it did this week. Technical difficulties aside, I am looking forward to finding new gene variants and understanding their significance. 

Mobile Lab and Internet Based Approaches to Learning about Cognition 

Yesterday, after completing some work in the lab, Dr. Schrauwen invited her team to Dr. Matt Huentelman’s lecture on Internet and Mobile Lab-Based Approaches to Study Cognition. Dr. Hunetelman, who studies the aging brain, aimed to develop a study that investigated how the brain changes over time. For the internet-based approach of this study, over 700,000 participants took the MindCrowd brain test, which involved paired-associate learning exams alongside a questionnaire on each subject’s demographic and health information. If you are interested in taking this short test and contributing to the study, please click here.  

The mobile-based approach took place in a mobile health lab, a horse trailer repurposed to accommodate a low-field MRI and other equipment to observe brain physiology and activity. Subjects in this mobile-based study participated in the following: 

  • MindCrowd brain game 
  • Scent detection activity
  • Grip strength testing

The investigators in this lab also collected blood pressure information, blood samples, MRI scans, and retinal images. The image produced by the low-field MRI scan was later completed using Artificial Intelligence and 3D printed for each participant to keep. Isn’t that a cool token to commemorate your participation in scientific research? 

Takeaways and Next Steps 

Tuning in to the lecture helped me realize the different approaches to data collection in neuroscience research. Although my project focuses on a niche sample, it is also important to understand how to make studies accessible. Although I will continue searching for genetic information in the Finnish families that the lab has information on, the broader implications for certain genetic differences must be kept in mind as well. 

Thanks for tuning in!

-Caitlin

More Posts

Comments:

All viewpoints are welcome but profane, threatening, disrespectful, or harassing comments will not be tolerated and are subject to moderation up to, and including, full deletion.

    tanay_n
    Caitlin, Dr. Hunetelman's study is fascinating! I'm curious, given the challenges with AI MARRVEL prioritizing intronic variants over exonic ones, how do you see the role of AI evolving in variant interpretation?
    Alana Rothschild
    Hi Caitlin! Wow- you are doing such great work. Your test that contributes to the study is very well done. It must have taken a lot of time and effort. Keep up the amazing work!
    allison_y
    Wonderful work Caitlin! I love how you’re combining cutting-edge AI tools with genetic research. It’s also awesome to see how you’re making the most of each new challenge and opportunity in your research. Can't wait to see the rest of your project!
    caitlin_e
    Thank you so much Allison! I can't wait to share more about my project in my final presentation!
    caitlin_e
    Thank you Ms. Rothschild! I appreciate it!
    caitlin_e
    Hi Tanay! AI MARRVEL is a prioritization tool, meaning that the software is trained to value certain factors over others while assessing data. That being said, some factors that are of little significance for diagnosing certain diseases are being amplified, while more important factors are overlooked. As a result, this interferes with the accuracy of the Artificial Intelligence. These Artificial Intelligence systems can improve by providing more accurate insight on genetic variants and their phenotypic implications. In addition, the Artificial Intelligence models can be trained on more specific information (instead of broad databases) when a lab is searching for a particular condition's genetic basis. Finally, effective communication with the system developers and the scientists using the AI system for genetic analysis is necessary in order to improve the program and apply it effectively.
    camille_bennett
    I absolutely love that you also explored access to studies this week! Great way to encourage participants, as well.
    Rahul Patel
    Hey Caitlin, great work diving into the genetic analysis! It’s interesting that ClinPrior is showing more promise than MARRVEL—can you share more on how ClinPrior outperforms the other tool in identifying gene variants? Also, the mobile lab approach to studying cognition sounds fascinating, especially with the 3D printed MRI images. Looking forward to seeing where your research goes next!
    aashi_h
    Caitlin, this is such a fascinating project! I wish I could have a 3D printed model like that. What are the broader implications for certain genetic differences that we must keep in mind as you explore Finnish genetics?
    caitlin_e
    Hi Rahul! I appreciate your enthusiasm surrounding mobile lab research. As for your question, part of the reason why ClinPrior outperforms other tools is because it places more of an emphasis on phenotype. In fact, the first step of the ClinPrior algorithm is calculating a Phenotypical Association Metric Calculation. Afterwards, a genetic link is explored further.
    caitlin_e
    Hi Aashi! I would like a 3D printed model of my brain as well. To answer your question, as we explore Finnish genetics, we must draw a distinction between the magnitude and scope of the study. The sample was small and niche, and there may be less genetic diversity in such a specific population. Despite this supposed limitation, a trend observed in this data set has the potential to be observed in other individuals from other areas as well. The findings of this study can be applied to broader contexts despite immense genetic differences in other parts of the world because genetic code is universal.
    caitlin_e
    Thank you Ms. Bennett! I'm happy to see so much involvement :)

Leave a Reply

Your email address will not be published. Required fields are marked *