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.

You might learn a lot about that house—the rooms, the doors, the furniture—but you would miss the bigger story: the streets, the neighborhoods, and all the ways the city changes from one place to another.

Something similar happens when scientists study bacterial genomes one at a time.

Over the past decade, researchers have collected thousands of bacterial genomes. These genomes contain an enormous amount of information about how bacteria survive, adapt, and evolve. But comparing thousands of individual genomes can quickly become a computational maze.

That is where PANORAMA enters the story.

Developed by Jérôme Arnoux and colleagues, PANORAMA is a computational tool designed to explore bacterial pangenomes—the complete collection of genetic possibilities found across a species or group of related organisms. Instead of looking at every genome as an isolated object, the approach represents their shared and variable genetic features as a graph.

Think of this graph as a giant subway map.

Some stations appear on almost every route. These represent genes that are highly conserved. Other stations appear only on certain routes, representing genes that some bacteria possess while others do not. By looking at the entire network, scientists can begin to see not just *what genes exist*, but how they are organized and how biological systems are distributed across populations.

The researchers put PANORAMA to the test using 941 genomes of Pseudomonas aeruginosa, a bacterium important in human health. They used the tool to investigate biological systems, including bacterial defense mechanisms against viruses known as bacteriophages. They then expanded the analysis to more than 6,000 genomes from four Enterobacteriaceae species.

The result is more than a faster way to process data.

PANORAMA allows researchers to ask a bigger question: What can an entire microbial species do, genetically speaking?

By comparing pangenomes, the researchers could identify systems shared between species as well as distinctive features. They could also find recurring genomic locations where genetic material is inserted—clues that may reveal common evolutionary processes.

And this is perhaps the most exciting part of the story.

Every bacterial genome is like a page in a huge evolutionary book. Until recently, reading thousands of those pages together was difficult. PANORAMA provides a way to turn those pages into a map, allowing scientists to see patterns that might disappear when each genome is studied separately.

The study, published in PLOS Computational Biology in July 2026, presents PANORAMA as a foundation for large-scale comparative pangenomics. The software and accompanying analysis resources are openly available, giving other researchers the opportunity to explore microbial diversity themselves.

So the story is not really about one bacterium or one genome.

It is about changing the way we look at life.

Instead of asking, “What is inside this genome?”, scientists can increasingly ask, “What is the full genetic landscape of this species—and how did it become this way?”

And sometimes, when you stop looking at one house and finally see the whole city, the most interesting discoveries are hiding in the streets between them.

Read more at https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1013856

A recent study by Webster and Chapman introduces Buscogeny (https://github.com/Jwebster89/Buscogeny), an open-source bioinformatics pipeline designed to make this process easier. The tool uses BUSCO (Benchmarking Universal Single-Copy Orthologs), which identifies conserved genes that are expected to occur as single copies in particular groups of organisms. These genes act like evolutionary landmarks, allowing researchers to compare genomes and investigate their relationships.

Buscogeny brings several steps of phylogenomic analysis into one workflow. It identifies BUSCO genes, extracts and aligns their sequences, removes unreliable regions, combines information from multiple genes, and uses the resulting data to construct a phylogenetic tree. It also provides quality-control information, helping researchers identify genomes that may be too incomplete for reliable analysis.

The authors demonstrate the usefulness of Buscogeny with bacterial and fungal genomes, including Alternaria. This is particularly valuable because fungal genomes can vary greatly in quality and annotation. Ultimately, Buscogeny helps transform massive amounts of genomic data into an understandable evolutionary story. Each DNA sequence becomes a clue, each conserved gene becomes a landmark, and together these clues reveal a picture of how life is connected.

More at https://link.springer.com/article/10.1007/s10123-025-00752-6?

WiseScaffolder is a stand-alone semi-automatic application for genome scaffolding of pre-assembled contigs using mate-pair data. It also produces editable scaffold maps, allowing either to build gapped scaffolds or usable as a common thread for the manual improvement of scaffolds.

Description 

WiseScaffolder includes 4 subcommands: dumpconfig generates a configuration file that notably specifies the average insert size of the mate-pair library preprocess allows the detection and correction of chimerae, the estimation of contigs copy number and produces valuable outputs for the manual improvement of scaffolds scaffold constitutes the central scaffold-builder and comprises two modules:

i) the interative_scaffold_extender, which works with big, unambiguous contigs, or when they run out, single copy contigs, and

ii) the small_contig_inserter, which inserts the small contigs within scaffolds buildfasta converts the scaffold(s) map(s) into Fasta sequences.

Address of the bookmark: http://abims.sb-roscoff.fr/wisescaffolder

Address of the bookmark: http://genome.cs.nthu.edu.tw/Multi-CAR/

Address of the bookmark: http://omega.omicsbio.org/

So far miniasm is in early development stage. It has only been tested on a dozen of PacBio and Oxford Nanopore (ONT) bacterial data sets. Including the mapping step, it takes about 3 minutes to assemble a bacterial genome. Under the default setting, miniasm assembles 9 out of 12 PacBio datasets and 3 out of 4 ONT datasets into a single contig. The 12 PacBio data sets are PacBio E. coli sample, ERS473430, ERS544009, ERS554120, ERS605484, ERS617393, ERS646601, ERS659581, ERS670327, ERS685285, ERS743109 and a deprecated PacBio E. coli data set. ONT data are acquired from the Loman Lab.

For a C. elegans PacBio data set (only 40X are used, not the whole dataset), miniasm finishes the assembly, including reads overlapping, in ~10 minutes with 16 CPUs. The total assembly size is 105Mb; the N50 is 1.94Mb. In comparison, the HGAP3produces a 104Mb assembly with N50 1.61Mb. This dotter plot gives a global view of the miniasm assembly (on the X axis) and the HGAP3 assembly (on Y). They are broadly comparable. Of course, the HGAP3 consensus sequences are much more accurate. In addition, on the whole data set (assembled in ~30 min), the miniasm N50 is reduced to 1.79Mb. Miniasm still needs improvements.

Miniasm confirms that at least for high-coverage bacterial genomes, it is possible to generate long contigs from raw PacBio or ONT reads without error correction. It also shows that minimap can be used as a read overlapper, even though it is probably not as sensitive as the more sophisticated overlapers such as MHAP and DALIGNER. Coupled with long-read error correctors and consensus tools, miniasm may also be useful to produce high-quality assemblies.

Minimap and miniasm are ultrafast tools for (i) mapping and (ii) assembly. Designed for long, noisy reads, they do not have a correction or consensus step, and therefore the resulting assemblies are contiguous (i.e. long) but very noisy (i.e. full of errors)

We start with an all against all comparison:

minimap -Sw5 -L100 -m0 -t8 reads.fq reads.fq | gzip -1 > reads.paf.gz

Then we can assemble

miniasm -f reads.fq reads.paf.gz > reads.gfa

Convert GFA to FASTA:

awk '/^S/{print ">"$2"\n"$3}' reads.gfa | fold > reads.fa

And then count how many contigs:

grep ">" reads.fa | wc -l

# Download sample PacBio from the PBcR website
wget -O- http://www.cbcb.umd.edu/software/PBcR/data/selfSampleData.tar.gz | tar zxf -
ln -s selfSampleData/pacbio_filtered.fastq reads.fq
# Install minimap and miniasm (requiring gcc and zlib)
git clone https://github.com/lh3/minimap && (cd minimap && make)
git clone https://github.com/lh3/miniasm && (cd miniasm && make)
# Overlap
minimap/minimap -Sw5 -L100 -m0 -t8 reads.fq reads.fq | gzip -1 > reads.paf.gz
# Layout
miniasm/miniasm -f reads.fq reads.paf.gz > reads.gfa

Address of the bookmark: https://github.com/lh3/miniasm

https://github.com/ggonnella/rgfa

Address of the bookmark: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5103826/

python finisherSC.py destinedFolder mummerPath

If you are running on server computer and would like to use multiple threads, then the following commands can generate 20 threads to run FinisherSC.

python finisherSC.py -par 20 destinedFolder mummerPath

Sometimes, if the names of raw reads and contigs consists of special characters/formats, FinisherSC/MUMmer may not parse them correctly. In that case, you want to have a quick renaming of the names of contigs/reads in contigs.fasta or raw_reads.fasta using the following command.

    perl -pe 's/>[^\$]*$/">Seg" . ++$n ."\n"/ge' raw_reads.fasta > newRaw_reads.fasta
    cp newRaw_reads.fasta raw_reads.fasta
    perl -pe 's/>[^\$]*$/">Seg" . ++$n ."\n"/ge' contigs.fasta > newContigs.fasta
    cp newContigs.fasta contigs.fasta

Address of the bookmark: https://github.com/kakitone/finishingTool