The Power of Evo 2: AI Meets DNA

Evo 2 is the largest AI model for biology ever created, trained on an astonishing 9.3 trillion DNA "letters" (nucleotides) carefully selected from genomes spanning the entire tree of life. This massive dataset ensures that Evo 2 can recognize patterns and relationships in genetic sequences at an unparalleled scale.

For the first time, scientists can design DNA with AI, moving beyond simple sequence analysis to active DNA generation. Evo 2 enables researchers to predict, modify, and even create entire genetic sequences, opening new possibilities in medicine, agriculture, and synthetic biology.

Decoding the Dark Genome

One of the biggest challenges in genetics is understanding the non-coding regions of DNA—vast stretches of the genome that do not code for proteins but play crucial roles in regulating gene expression. These regions control when and how genes are activated, influencing everything from development to disease.

Evo 2 is designed to decode these non-coding elements, helping researchers uncover their functions and use this knowledge to develop gene-based therapies, synthetic life forms, and precision agriculture solutions.

From Reading DNA to Writing It

To put Evo 2’s impact into perspective:

• Previous AI models could "read" DNA like a book, analyzing genetic sequences and identifying patterns.
• Evo 2 can "write" entirely new DNA, designing functional genes, chromosomes, and even full genomes from scratch.

This means scientists can now engineer biological systems with AI, designing new proteins, metabolic pathways, and genetic circuits to address real-world challenges.

A Step Toward Generative Biology

The Arc Institute describes Evo 2 as a major step toward "generative biology"—a revolutionary approach where AI is used to create novel biological structures rather than just analyzing existing ones. This could lead to breakthroughs such as:

• New medicines: AI-generated enzymes and proteins tailored for targeted therapies.
• Disease-resistant crops: Genetically optimized plants for higher yield and climate resilience.
• Synthetic organisms: Custom-designed microbes for bioremediation, biofuel production, and industrial applications.

An Open-Source Revolution

Unlike many proprietary AI models, Evo 2 is open source, making its capabilities accessible to researchers worldwide. This democratization of AI-driven biology means that scientists from different disciplines can collaborate, experiment, and innovate, accelerating discoveries in genetic engineering and synthetic biology.

With Evo 2, the boundaries of what’s possible in DNA design, genetic engineering, and biological innovation are being redrawn. The future of life sciences is no longer just about understanding life’s code—it’s about writing it.

That is the challenge researchers face when studying how bacteria adapt to different environments. A bacterial gene may appear to be linked to a particular habitat, when the real reason is simply that closely related bacteria happen to live there.

In their 2025 Genome Biology paper, Bujdoš, Walter, and O’Toole introduce aurora, a machine-learning tool designed to tackle this problem.

Aurora identifies potentially mislabeled or unusual bacterial strains before performing genome-wide association studies (GWAS). By cleaning up the dataset and accounting for bacterial evolutionary relationships, it can help researchers find genetic features that are more genuinely connected to habitat adaptation.

The researchers tested aurora using simulated and real bacterial datasets and found that it could recover important genetic associations even when datasets contained misleading labels.

The bigger lesson is simple: better biological discoveries often begin with better data.

Aurora gives researchers a new way to separate real genetic clues from misleading ones—and could help us better understand how bacteria adapt, survive, and evolve in the environments they call home.

Based on Bujdoš et al., “aurora: a machine learning GWAS tool for analyzing microbial habitat adaptation,” Genome Biology (2025).[Read the original paper](https://link.springer.com/article/10.1186/s13059-025-03524-7?)

Research Directions

Stony corals: evolution of novelty and photosymbiosis

Symbiotic acoels: cell type evolution and photosymbiosis

Animal genomes: structural evolution and gene regulation

https://sgel.biodiv.tw/home

Tivicay is indicated for infected adults, both treatment-naive and treatment-experienced, including those treated with other integrase strand transfer inhibitors; as well as for children aged 12 years and older weighing at least 40 kg (approx. 88 lbs). The drug - which was used without a pharmacokinetic boosting agent - interferes with one of the enzymes necessary for HIV to multiply, and can be taken with or without food and at any time of day.

http://www.fda.gov/ForConsumers/ByAudience/ForPatientAdvocates/HIVandAIDSActivities/ucm364869.htm

Link @ http://www.bioinf.uni-freiburg.de/

Lab page @ http://graveleylab.cam.uchc.edu/Graveley/index.html

We work in a highly interdisciplinary environment at the interface of Computer Science and Biology. Since its inception, our lab has eagerly engaged in collaborative research partnerships with biological and experimental collaborators, facilitated by our affiliation with the Broad Institute and the Computational and Systems Biology initiative (CSBi) at MIT, our participation in the Epigenome Roadmap, ENCODE, and modENCODE consortia, and by several other ongoing collaborations at MIT, Harvard, and the Harvard Medical School affiliated hospitals.

http://compbio.mit.edu/

The bpRNA code is written in perl and requires the Graph perl module. Several additional scripts for analysis are included. The source code is available at http://github.com/hendrixlab/bpRNA.

Address of the bookmark: http://github.com/hendrixlab/bpRNA

Protein-protein and protein-DNA/RNA docking based on a hybrid algorithm of template-based modeling and ab initio free docking.

The HDOCK server distinguishes itself from similar docking servers in its ability to support amino acid sequences as input and a hybrid docking strategy in which experimental information about the protein–protein binding site and small-angle X-ray scattering can be incorporated during the docking and post-docking processes.

Address of the bookmark: http://hdock.phys.hust.edu.cn/

Why Normalize RNA-Seq Data?

Normalization is a crucial step in RNA-Seq analysis for the following reasons:

• Sequencing depth: Different RNA-Seq experiments produce varying numbers of reads, making direct comparisons between samples misleading.

• Gene length: Longer genes inherently generate more reads, irrespective of their actual expression level.

• Bias reduction: Normalization mitigates technical biases, enabling meaningful biological interpretation.

TPM (Transcripts Per Million)

TPM measures the proportion of reads mapped to a transcript, normalized by transcript length and sequencing depth. It is calculated as:

Key Features:

1. Proportionality: TPM values sum to 1,000,000 across all transcripts in a sample, making it easier to compare between samples.

2. Intuitive interpretation: TPM values directly represent the abundance of transcripts in a sample.

3. Preferred for comparisons: TPM facilitates between-sample comparisons better than FPKM.

FPKM (Fragments Per Kilobase Million)

FPKM normalizes read counts by transcript length and sequencing depth, but without enforcing proportionality like TPM. It is defined as:

Key Features:

1. Historical significance: FPKM was one of the first normalization methods used for RNA-Seq.

2. Single-end vs. paired-end: In paired-end sequencing, FPKM becomes RPKM (Reads Per Kilobase Million).

3. Limited utility: FPKM values are not as robust as TPM for cross-sample comparisons due to lack of proportionality.

CPM (Counts Per Million)

CPM normalizes raw read counts by sequencing depth, without considering gene length. It is expressed as:

Key Features:

1. Simplicity: CPM is straightforward and computationally less intensive.

2. Application: Suitable for non-length-dependent analyses, such as comparing total expression levels or differential expression analysis.

3. Gene length agnostic: CPM does not correct for gene length, making it less ideal for measuring expression levels.

When to Use Each Method

• TPM: Best for comparing expression levels between samples, especially when transcript length and sequencing depth vary.

• FPKM: Useful for historical consistency but generally replaced by TPM.

• CPM: Ideal for differential expression analysis when gene length normalization is unnecessary.

Conclusion

Choosing the right normalization method depends on the specific objectives of your RNA-Seq analysis. TPM’s proportionality and robustness make it the preferred choice for most applications, while CPM serves well for differential expression studies. Although FPKM paved the way for RNA-Seq normalization, it has largely been supplanted by TPM in modern workflows. Understanding these methods and their nuances ensures accurate and meaningful interpretations of RNA-Seq data.

References:

1. Li, B., & Dewey, C. N. (2011). RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics.

2. Trapnell, C., et al. (2010). Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nature Biotechnology.

3. Law, C. W., et al. (2014). voom: precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biology.