🧬 AlphaFold 3 predicts the structure and interactions of all of life’s molecules


▪Blog: https://blog.google/technology/ai/google-deepmind-isomorphic-alphafold-3-ai-model/
▪Nature: https://www.nature.com/articles/s41586-024-07487-w
▪Two Minute Papers: https://www.youtube.com/watch?v=Mz7Qp73lj9o

@BioinformaticsA

Free 4-week long course for beginners:

What you'll learn

Basics of R
Basics of Python
How to analyze bulk RNAseq count data
How to analyze single cell RNAseq count data

https://www.coursera.org/learn/fundamental-skills-in-bioinformatics

History and Evolution of Protein Structure Prediction

This lecture introduces the fascinating history of protein structure prediction, highlighting the significant advancements in this field.

Early Stages:

- 1950s: Fred Sanger revolutionized protein sequencing, enabling the comparison of sequences to identify similarities.
- 1970s: Protein Data Bank (PDB) was established to store protein structures, laying the foundation for structural comparisons.

Advancements in Computational Methods:

- 1990s: Algorithms like BLAST emerge for rapid sequence searching.
- 1990s: Computational approaches for predicting protein structures are proposed, but accuracy suffers.
- 1994: Critical Assessment of Structure Prediction (CASP) competition is established to assess protein structure prediction algorithms objectively.

Recent Progress:

- 2012: Deep learning techniques like HHpred and Rosetta emerge as highly accurate predictors in CASP competition.
- 2018: AlphaFold from DeepMind utilizes deep learning to learn the rules from existing sequences and structures, achieving unprecedented accuracy.
- 2020: AlphaFold2 further improves accuracy, suggesting that the protein structure prediction problem is largely solved.

Unresolved Challenges:

- Predicting protein-protein interactions in complex systems.
- Predicting structures of large molecules.

Future Directions:

- Continued development of AI-powered algorithms to tackle more complex protein structures.
- Integration of structural predictions with other omics data like proteomics and metabolomics.

Conclusion:

The field of protein structure prediction has witnessed remarkable progress, driven by technological advancements and the power of AI. While significant challenges remain, the future looks promising, potentially revolutionizing our understanding of proteins and their functions.

STAT115 Chapter 2.1 Protein Wave

#LLMs
@BioinformaticsA

The Second Wave of Bioinformatics - Gene Expression Wave

The second wave of bioinformatics, the gene expression wave, emerged in the 1970s with the invention of Northern blot, a technique to measure gene expression. Microarrays revolutionized this further, enabling the simultaneous measurement of hundreds to thousands of genes. Commercial microarray platforms emerged with increased reproducibility and affordability.

Key advancements include:

- Microarrays:

- Ability to measure expression of hundreds to thousands of genes at once.
- Applications in disease diagnosis, cancer research, and drug development.

Commercial Microarrays:

- Increased reproducibility and affordability compared to hand-spotted arrays.

Single-Cell Gene Expression:

- Technologies like RNA-Seq can capture gene expression data from individual cells, revealing cellular heterogeneity within tissues.

Applications:

- Precise diagnosis of diseases like leukemia.
- Identification of genes associated with disease progression.
- Drug development and personalized medicine.
- Understanding cell behavior and tissue development.

Future Directions:

- Affordable and accessible technologies like RazorSeq and TranscriptomeSeq enable large-scale gene expression studies.
- Computational analysis of massive datasets is crucial for understanding complex biological phenomena.

STAT115 Chapter 2.2 Expression Wave

#LLMs
#Summary
@BioinformaticsA

📄 Deep Generative Models for Therapeutic Peptide Discovery: A Comprehensive Review

📗 Journal: ACM Computing Surveys (🔥I.F.=23.8)
🗓 Publish year: 2025

🧑‍💻Authors: Leshan Lai, Yuansheng Liu, Bosheng Song, ...
🏢Universities: Hunan University, China - State University of New York, USA

📎 Study the paper

📲Channel: @Bioinformatics
#review #llm #peptide #therapeutic

!pip install -q bioservices

from bioservices import UniProt

u = UniProt()
results = u.search("83333")

import pandas as pd
from io import StringIO

df = pd.read_csv(StringIO(results), sep="\t")
df.head()

Escherichia coli (taxon ID: 83333)

@BioinformaticsA

If you are not familiar with Python, a great place to start is Learn Python, where you will find many tutorials on the basics of the language.

@BioinformaticsA