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

A recent study published in Genome Biology (https://link.springer.com/article/10.1186/s13059-026-04280-y) addressed this challenge by introducing dicast (https://github.com/burgshrimps/dicast), a machine-learning approach for detecting structural variants in short-read sequencing data. Structural variants, such as large deletions, insertions, duplications, and inversions, can have significant biological and clinical implications, yet they are challenging to identify with short-read technologies. Because different detection methods frequently yield divergent results, researchers commonly employ consensus calling, considering a variant valid only if multiple tools detect it. While this approach reduces false positives, it relies on the potentially flawed assumption that the majority is always correct.

The researchers explored the impact of evaluating the supporting evidence for each variant, rather than simply tallying the number of algorithms that identified it. To establish a ground truth, they analyzed nine genomes using multiple sequencing technologies and 15 detection methods, initially identifying approximately 35 million potential variants. Through extensive filtering, evidence integration, and manual review of over 11,500 variants, they developed a robust benchmark comprising more than 236,000 structural variants. The findings underscored the complexity of the problem: short-read methods detected fewer than half of deletions and less than 10 percent of insertions, with performance declining markedly in repetitive genomic regions. In contrast, long-read technologies demonstrated superior detection capabilities. However, replacing the substantial volume of existing short-read data in clinical and research settings is not immediately feasible. Consequently, the researchers questioned whether short-read data might harbor more information than conventional analytical pipelines currently extract.

This line of inquiry led to the development of dicast. Rather than merely confirming agreement among multiple tools, dicast identifies patterns in sequencing data, including split and clipped reads, discordant read pairs, alignment characteristics, and the surrounding genomic context. An XGBoost machine-learning model evaluates which combinations of these signals are indicative of genuine structural variants. Thus, the approach shifts from tallying algorithmic consensus to interpreting the underlying evidence.

The researchers subsequently conducted a targeted evaluation by examining structural variants detected by only a single short-read tool, which are typically missed by consensus-based approaches. dicast successfully recovered approximately 81% of these single-caller deletions, insertions, and duplications. The signals for these variants were present in the data, but conventional filtering methods failed to integrate them effectively.

The utility of dicast was further demonstrated in cohorts with rare diseases, including congenital limb malformations, atrial fibrillation, and neuromuscular disorders. In one instance, dicast achieved a deletion recall rate of approximately 0.96, compared to 0.74 using consensus calling. The median number of variants requiring manual review was 29 per sample. Among 31 experimentally validated variants that standard filters would have missed, dicast identified 12, whereas consensus calling detected only one.

Overall, dicast identified approximately 20 percent more potential disease-causing deletions than consensus-based methods. While a 20 percent increase may appear modest, in clinical genomics such improvements can have significant practical implications. Missing a deletion may leave a case unresolved, whereas detecting a structural variant can provide critical diagnostic insights.

The study does not claim that machine learning has rendered short-read sequencing superior to long-read approaches. Instead, the results underscore the effectiveness of long-read sequencing for structural variant detection. However, dicast highlights a more nuanced perspective: substantial biological information may still be recoverable from the extensive short-read datasets already available.

The principal lesson extends beyond the detection of structural variants. For many years, bioinformatics pipelines have relied on threshold-based criteria, such as minimum coverage, quality scores, or support from multiple tools. While these rules are useful, biological phenomena do not always conform to rigid checklists; multiple weak signals, when considered collectively, can provide compelling evidence.

This perspective prompts consideration of the solitary variant: one algorithm identifies it, while several others do not. Traditional consensus techniques might have dismissed it, yet machine learning approaches evaluate the available evidence to determine whether the variant is plausible.

Occasionally, the most significant variant within a genome is the one that is almost universally overlooked.

Read more at https://link.springer.com/article/10.1186/s13059-026-04280-y

Tel +49 451 3101 8301
cardiogenetics(at)uni-luebeck.de

http://cardiogenetics-luebeck.de/

The biggest and perhaps most fearsome of the world's big cats, the tiger, shares 95.6 percent of its DNA with humans' cute and furry companions, domestic cats.

The new research showed that big cats have genetic mutations that enabled them to be carnivores. The team also identified mutations that allow snow leopards to thrive at high altitudes.

Reference:

http://www.nbcnews.com/science/your-cat-ferocious-tigers-share-lot-95-6-percent-their-4B11182690

http://timesofindia.indiatimes.com/home/environment/flora-fauna/Gene-mapping-of-tiger-completed/articleshow/22671681.cms

Paper:

http://www.nature.com/ncomms/2013/130917/ncomms3433/full/ncomms3433.html

http://en.wikipedia.org/wiki/Sequence_assembly 

https://genomebiology.biomedcentral.com/articles/10.1186/s13059-016-0951-y

Address of the bookmark: https://sourceforge.net/projects/operasf/

Again, running spades.py will show you the options:

spades.py

This produces:

SPAdes genome assembler v3.10.1

Usage: /usr/local/SPAdes-3.10.1-Linux/bin/spades.py [options] -o <output_dir>

Basic options:
-o      <output_dir>    directory to store all the resulting files (required)
--sc                    this flag is required for MDA (single-cell) data
--meta                  this flag is required for metagenomic sample data
--rna                   this flag is required for RNA-Seq data
--plasmid               runs plasmidSPAdes pipeline for plasmid detection
--iontorrent            this flag is required for IonTorrent data
--test                  runs SPAdes on toy dataset
-h/--help               prints this usage message
-v/--version            prints version

Input data:
--12    <filename>      file with interlaced forward and reverse paired-end reads
-1      <filename>      file with forward paired-end reads
-2      <filename>      file with reverse paired-end reads
-s      <filename>      file with unpaired reads
--pe<#>-12      <filename>      file with interlaced reads for paired-end library number <#> (<#> = 1,2,..,9)
--pe<#>-1       <filename>      file with forward reads for paired-end library number <#> (<#> = 1,2,..,9)
--pe<#>-2       <filename>      file with reverse reads for paired-end library number <#> (<#> = 1,2,..,9)
--pe<#>-s       <filename>      file with unpaired reads for paired-end library number <#> (<#> = 1,2,..,9)
--pe<#>-<or>    orientation of reads for paired-end library number <#> (<#> = 1,2,..,9; <or> = fr, rf, ff)
--s<#>          <filename>      file with unpaired reads for single reads library number <#> (<#> = 1,2,..,9)
--mp<#>-12      <filename>      file with interlaced reads for mate-pair library number <#> (<#> = 1,2,..,9)
--mp<#>-1       <filename>      file with forward reads for mate-pair library number <#> (<#> = 1,2,..,9)
--mp<#>-2       <filename>      file with reverse reads for mate-pair library number <#> (<#> = 1,2,..,9)
--mp<#>-s       <filename>      file with unpaired reads for mate-pair library number <#> (<#> = 1,2,..,9)
--mp<#>-<or>    orientation of reads for mate-pair library number <#> (<#> = 1,2,..,9; <or> = fr, rf, ff)
--hqmp<#>-12    <filename>      file with interlaced reads for high-quality mate-pair library number <#> (<#> = 1,2,..,9)
--hqmp<#>-1     <filename>      file with forward reads for high-quality mate-pair library number <#> (<#> = 1,2,..,9)
--hqmp<#>-2     <filename>      file with reverse reads for high-quality mate-pair library number <#> (<#> = 1,2,..,9)
--hqmp<#>-s     <filename>      file with unpaired reads for high-quality mate-pair library number <#> (<#> = 1,2,..,9)
--hqmp<#>-<or>  orientation of reads for high-quality mate-pair library number <#> (<#> = 1,2,..,9; <or> = fr, rf, ff)
--nxmate<#>-1   <filename>      file with forward reads for Lucigen NxMate library number <#> (<#> = 1,2,..,9)
--nxmate<#>-2   <filename>      file with reverse reads for Lucigen NxMate library number <#> (<#> = 1,2,..,9)
--sanger        <filename>      file with Sanger reads
--pacbio        <filename>      file with PacBio reads
--nanopore      <filename>      file with Nanopore reads
--tslr  <filename>      file with TSLR-contigs
--trusted-contigs       <filename>      file with trusted contigs
--untrusted-contigs     <filename>      file with untrusted contigs

Pipeline options:
--only-error-correction runs only read error correction (without assembling)
--only-assembler        runs only assembling (without read error correction)
--careful               tries to reduce number of mismatches and short indels
--continue              continue run from the last available check-point
--restart-from  <cp>    restart run with updated options and from the specified check-point ('ec', 'as', 'k<int>', 'mc')
--disable-gzip-output   forces error correction not to compress the corrected reads
--disable-rr            disables repeat resolution stage of assembling

Advanced options:
--dataset       <filename>      file with dataset description in YAML format
-t/--threads    <int>           number of threads
                                [default: 16]
-m/--memory     <int>           RAM limit for SPAdes in Gb (terminates if exceeded)
                                [default: 250]
--tmp-dir       <dirname>       directory for temporary files
                                [default: <output_dir>/tmp]
-k              <int,int,...>   comma-separated list of k-mer sizes (must be odd and
                                less than 128) [default: 'auto']
--cov-cutoff    <float>         coverage cutoff value (a positive float number, or 'auto', or 'off') [default: 'off']
--phred-offset  <33 or 64>      PHRED quality offset in the input reads (33 or 64)
                                [default: auto-detect]

As you can see this is also a “pipeline” of tools that can be switched on or off. SPAdes takes quite a long time, so for the purposes of this practical, something like this may suffice:

spades.py -t 4 \
          -m 32 \
          -k 31,51,71 \
          --only-assembler \
          -1 miseq.1.fastq -2 miseq.2.fastq \
          --nanopore minion.fastq \
          -o hybrid_assembly

In turn, these parameters mean

• use 4 threads
• max memory is 32Gb
• use 3 kmer values to build the de bruijn graph(s) - 31, 51 and 71
• only run the assembler, not the correction algorithm (for speed)
• read 1 and read 2 of the MiSeq data
• the nanopore data
• put the output in folder “hybrid_assembly”

Address of the bookmark: ftp://ftp.genomics.org.cn/pub/cope

Prokka – Standalone command line tool, takes just a few minutes per genome. This is the best way to get good quality annotation in a flash, which is particularly useful if you have loads of genomes or need to annotate a pangenome or metagenome. Note however that the quality of functional information is not as good as RAST, and you will need several extra steps if you want to do functional profiling and pathway analysis of your genome(s)… which is in-built in RAST.

NCBI Prokaryotic Genome Annotation Pipeline is designed to annotate bacterial and archaeal genomes (chromosomes and plasmids).

Genome annotation is a multi-level process that includes prediction of protein-coding genes, as well as other functional genome units such as structural RNAs, tRNAs, small RNAs, pseudogenes, control regions, direct and inverted repeats, insertion sequences, transposons and other mobile elements.

PGAP: NCBI has developed an automatic prokaryotic genome annotation pipeline that combines ab initio gene prediction algorithms with homology based methods. The first version of NCBI Prokaryotic Genome Automatic Annotation Pipeline (PGAAP; see Pubmed Article) developed in 2005 has been replaced with an upgraded version that is capable of processing a larger data volume.  NCBI's annotation pipeline depends on several internal databases and is not currently available for download or use outside of the NCBI environment.

BEACON (automated tool for Bacterial GEnome Annotation ComparisON), a fast tool for an automated and a systematic comparison of different annotations of single genomes. The extended annotation assigns putative functions to many genes with unknown functions. BEACON is available under GNU General Public License version 3.0 and is accessible at: http://www.cbrc.kaust.edu.sa/BEACON/.

BlastKOLA: Assigns K numbers to the user's sequence data by BLAST searches, respectively, against a nonredundant set of KEGG GENES. KOALA (KEGG Orthology And Links Annotation) is KEGG's internal annotation tool for K number assignment of KEGG GENES using SSEARCH computation. Annotate Sequence in KEGG Mapper and Pathogen Checker in KEGG Pathogen are special interfaces to this server and can be executed in an interactive mode. BlastKOALA is suitable for annotating fully sequenced genomes.

PAGIT: Provides a toolkit for improving the quality of genome assemblies created via an assembly software. PAGIT compiled four tools: (i) ABACAS which classifies and orientates contigs and estimates the sizes of gaps between them; (ii) IMAGE uses paired-end reads to extend contigs and close gaps within the scaffolds; (iii) ICORN for identifying and correcting small errors in consensus sequences and; (iv) RATT for help annotation. The software was mainly created to analyze parasite genomes of up to about 300 Mb.

MAKER: A portable and easily configurable genome annotation pipeline. MAKER allows smaller eukaryotic and prokaryotic genome projects to independently annotate their genomes and to create genome databases. It identifies repeats, aligns ESTs and proteins to a genome, produces ab-initio gene predictions and automatically synthesizes these data into gene annotations having evidence-based quality values. MAKER's inputs are minimal and its ouputs can be directly loaded into a Generic Model Organism Database (GMOD). They can also be viewed in the Apollo genome browser; this feature of MAKER provides an easy means to annotate, view and edit individual contigs and BACs without the overhead of a database. MAKER is available for download and can be tested online via the MAKER Web Annotation Service (MWAS).

MyPro is a software pipeline for high-quality prokaryotic genome assembly and annotation. It was validated on 18 oral streptococcal strains to produce submission-ready, annotated draft genomes. MyPro installed as a virtual machine and supported by updated databases will enable biologists to perform quality prokaryotic genome assembly and annotation with ease.

Magic-BLAST incorporates within the NCBI BLAST code framework ideas developed in the NCBI Magic pipeline, in particular hit extensions by local walk and jump (http://www.ncbi.nlm.nih.gov/pubmed/26109056), and recursive clipping of mismatches near the edges of the reads, which avoids accumulating artefactual mismatches near splice sites and is needed to distinguish short indels from substitutions near the edges.

Address of the bookmark: https://ncbi.github.io/magicblast/