Autism Spectrum Disorder Genome ML & Software Analysis

Autism spectrum disorder is a common medical condition referred to lack of brain development. The goal is to make a machine learning algorithm that can analyze autism and inform users about it.

Comment

Inspiration

Autism is well-known neurodevelopment disorder commonly found in adults and children which is characterized by inflexible behaviors, interests and can affect one's social life. Our project aimed to simulate a machine learning algorithm to analyze a specific DNA sequence from autism spectrum disorder and to allow the user to answer a questionnaire to determine if they likely to have autism using questions and their DNA sequence. While there are various DNA sequences involved with Autism. Our project focused on using Gene ID: NM_030627.4. The gene NM_030627.4 is known for coding the protein which is commonly known as CPEB4 and medically referred to as Idiopathic Autism (which is 80% of autistic patients), this protein leads to less neuronal activity, social awkwardness, sensitive to loud noises, and various symptoms that are well-recognized by medical professionals. While many people are aware of Autism's effects on disability due to famous shows such as the 'Good Doctor', what many people don't realize is that autism is on the rise as supported by the (CDC) Center of Disease Control Research. A study that was published in the JAMA Network Open found that children and adults diagnosed with autism spectrum disorder (ASD) increased by 175% over a decade and there could be various reasons such environmental factors like pollution, lack of healthcare, rise of media, The problem of Autism is that it's hard to identify and requires time and observational techniques. What's worse is that it is affecting many younger generations like children whose learning abilities shouldn't be hindered by Autism's effects.

The goal and inspiration of our team is to use technology such as Ai, Blockchain coding, and software development to create a efficient solution to model and track autism using bioinformatics and software development tools to spread awareness. By spreading awareness, it has led to more individuals to seek diagnoses and fund research to prevent its rise. While our model is not 100% accurate and we don't recommend it to be a final diagnosis, it does bring about the awareness on how autism can be analyzed and helps it make more identifiable.

What it does:

By asking users to do a questionnaire about their symptoms. Also, our code aims to get protein/DNA sequences and compare them using hamming distance methods to detect similarities with the proteins found in DNA affected by Autism versus a normal sequence. While our code doesn't predict "autism", it does show that maybe we take more into spreading awareness and show appreciation for what the healthcare and technician professionals can do as team to help people around the globe.

How we built it:

Using various software like National Center of Biotechnology Information Database, the gene ID for CPEB4 was stored and through different functions, a unique server port in Google Collab, the DNA sequence of the gene was translated into proteins and amino acids which could be used for analysis by Bioinformaticians and Biological scientists. Also, the code saved the longest protein as it has the most impact in the DNA sequence and compares it with the DNA sequence from the Software App. The Software App integrated with various web development and machine learning properties such as Flask were useful in making a questionnaire and allowing the user to make a decision between yes and no between symptoms. Depending on answer choices, they were given a response demonstrating if there was sufficient evidence for Autism-like symptoms.

Challenges we ran into

Some of the challenges our team ran into while coding is debugging, initial planning and finding a way to integrate all the code that we been trying to accomplish and building a story that can allow our code to function no matter the user that is involved.

Accomplishments that we're proud of:

Full genome analysis of Autism Protein using Gene DNA sequence from NCBI Database, the creation of software app to help patients and people detect signs of autism through awareness and questionnaire. • Programming logic to analyze questionnaire responses, determining the probability of autism based on the number of affirmative answers and providing personalized user feedback. • Learning to fix patches and errors that code was facing • Help Added a Database to store submission standards for connectivity for user for 1:1 relationship • Set up routes and branches in Python and Flask • Made app/browser to get request AHD requests, verifying logics to verify input.

What we learned:

Our team was working together as a first hackathon and it was a great experience working with software and connecting knowledge on python3, visual, and springbot experience. The team learned to collaborate and work together on a project that is more than just solving a problem, but an ability to save lives with code and artificial intelligence as it was useful in the creation of this project.

Also, we learned how to play with various file types like .ipynb and embed them with files related to our software which included python integrated with flask. Techniques of branching and demo of application work were learned.

What's next for Autism Prediction & Symptom Analysis using Machine Learning:

The next steps for Autism Prediction & Symptoms are use more algorithms and integrate more web searching in the code so it can allow the user to see a model that is improving and eventually analyze minor proteins in the Autism. Also to improve software design through more block chain coding and enhancement of web features.

Other minor issues have been integrating various coding files which was an issue that our team faced, in that case using more advanced and similar software can help our product and team to advance and fix lag issues.

Built With

Share this project:

Updates