Editing
Ai Genomics
(section)
Jump to navigation
Jump to search
Warning:
You are not logged in. Your IP address will be publicly visible if you make any edits. If you
log in
or
create an account
, your edits will be attributed to your username, along with other benefits.
Anti-spam check. Do
not
fill this in!
== <span style="color: #FFFFFF;">Understanding</span> == Genomics is inherently a machine learning problem: biological sequences are discrete strings over small alphabets (A, T, G, C for DNA; 20 amino acids for proteins), and the relationship between sequence and function is complex, non-linear, and learned from evolutionary history across billions of years. '''The protein folding revolution''': For 50 years, predicting a protein's 3D structure from its amino acid sequence was considered one of biology's grand challenges. DeepMind's AlphaFold2 (2020) solved this with near-experimental accuracy, using a combination of multiple sequence alignment, equivariant attention networks, and self-supervised learning on known protein structures. AlphaFold has predicted structures for virtually all known proteins β ~200M structures β in the AlphaFold Protein Structure Database, transforming drug discovery. '''Genomic language models''': BERT-style transformers pre-trained on DNA sequences learn rich representations of genomic function. Models like Enformer predict gene expression from DNA sequence by learning the regulatory grammar encoded in non-coding regions. DNABERT, Nucleotide Transformer, and Evo (trained on billions of DNA sequences) have achieved state-of-the-art on diverse genomic prediction tasks. '''Single-cell AI''': scRNA-seq measures gene expression in individual cells, generating sparse high-dimensional count matrices (cells Γ genes). AI tools like Seurat and Scanpy cluster cells by type; foundation models for single-cell data (scGPT, Geneformer) enable zero-shot cell type annotation, perturbation prediction, and drug response modeling. '''Polygenic Risk Scores (PRS)''': Aggregating thousands of small-effect genetic variants into a single disease risk score. Modern PRS methods use penalized regression (LASSO) and Bayesian approaches on GWAS summary statistics. PRS can predict cardiovascular disease, type 2 diabetes, and schizophrenia risk years before clinical onset. </div> <div style="background-color: #8B0000; color: #FFFFFF; padding: 20px; border-radius: 8px; margin-bottom: 15px;">
Summary:
Please note that all contributions to BloomWiki may be edited, altered, or removed by other contributors. If you do not want your writing to be edited mercilessly, then do not submit it here.
You are also promising us that you wrote this yourself, or copied it from a public domain or similar free resource (see
BloomWiki:Copyrights
for details).
Do not submit copyrighted work without permission!
Cancel
Editing help
(opens in new window)
Navigation menu
Personal tools
Not logged in
Talk
Contributions
Create account
Log in
Namespaces
Page
Discussion
English
Views
Read
Edit
View history
More
Search
Navigation
Main page
Recent changes
Random page
Help about MediaWiki
Tools
What links here
Related changes
Special pages
Page information