A review of machine learning methods for cancer characterization from microbiome data

npj Precision Oncology, Год журнала: 2024, Номер 8(1)

Опубликована: Май 30, 2024

Abstract Recent studies have shown that the microbiome can impact cancer development, progression, and response to therapies suggesting microbiome-based approaches for characterization. As cancer-related signatures are complex implicate many taxa, their discovery often requires Machine Learning approaches. This review discusses methods characterization from data. It focuses on implications of choices undertaken during sample collection, feature selection pre-processing. also ML model selection, guiding how choose an model, validation. Finally, it enumerates current limitations these may be surpassed. Proposed methods, based Random Forests, show promising results, however insufficient widespread clinical usage. Studies report conflicting results mainly due models with poor generalizability. We expect evaluating expanded, hold-out datasets, removing technical artifacts, exploring representations other than taxonomical profiles, leveraging advances in deep learning, developing better adapted characteristics data will improve performance generalizability enable usage clinic.

Язык: Английский

---

An oral microbiota-based deep neural network model for risk stratification and prognosis prediction in gastric cancer

Journal of Oral Microbiology, Год журнала: 2025, Номер 17(1)

Опубликована: Янв. 17, 2025

Background This study aims to develop an oral microbiota-based model for gastric cancer (GC) risk stratification and prognosis prediction.

Язык: Английский

---

Dysbiosis and extraintestinal cancers

Journal of Experimental & Clinical Cancer Research, Год журнала: 2025, Номер 44(1)

Опубликована: Фев. 7, 2025

Abstract The gut microbiota plays a crucial role in safeguarding host health and driving the progression of intestinal diseases. Despite recent advances remarkable correlation between dysbiosis extraintestinal cancers, underlying mechanisms are yet to be fully elucidated. Pathogenic microbiota, along with their metabolites, can undermine integrity barrier through inflammatory or metabolic pathways, leading increased permeability translocation pathogens. dissemination pathogens circulation may contribute establishment an immune-suppressive environment that promotes carcinogenesis organs either directly indirectly. oncogenic cascade always engages disruption hormonal regulation responses, induction genomic instability mutations, dysregulation adult stem cell proliferation. This review aims comprehensively summarize existing evidence points potential malignant transformation such as liver, breast, lung, pancreas. Additionally, we delve into limitations inherent current methodologies, particularly challenges associated differentiating low loads gut-derived microbiome within tumors from sample contamination symbiotic microorganisms. Although still controversial, understanding contribution translocated metabolites pathological continuum chronic inflammation could offer novel foundation for development targeted therapeutics.

Язык: Английский

---

Retraction Note: Microbiome analyses of blood and tissues suggest cancer diagnostic approach

Nature, Год журнала: 2024, Номер 631(8021), С. 694 - 694

Опубликована: Июнь 26, 2024

Язык: Английский

---

Community assembly modeling of microbial evolution within Barrett's esophagus and esophageal adenocarcinoma

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2025, Номер unknown

Опубликована: Янв. 16, 2025

Abstract Mathematical modeling of somatic evolution, a process impacting both host cells and microbial communities in the human body, can capture important dynamics driving carcinogenesis. Here we considered models for esophageal adenocarcinoma (EAC), cancer that has dramatically increased incidence over past few decades Western populations, with high case fatality rates due to late-stage diagnoses. Despite advancements genomic analyses precursor Barrett’s esophagus (BE), prevention EAC remains significant clinical challenge. Previous microbiome studies BE have focused on quantifying static abundance differences rather than evolutionary dynamics. Using whole genome sequencing data from tissues, first applied robust bioinformatics pipeline extract non-host DNA reads, mapped these putative reads taxa, retained those taxa coverage. When applying mathematical evolution sequential stages progression EAC, observed evidence neutral community assembly within normal tissue BE, but not EAC. In case-control study patients who progressed outcomes (CO) versus had non-cancer (NCO) during follow-up (mean=10.5 years), found Helicobacter pylori deviated significantly expectation NCO, suggesting factors related H. or infection itself may influence risk. Additionally, simulations incorporating selection recapitulated non-neutral behaviors datasets. Formally holds promise applications by offering deeper understanding involvement development.

Язык: Английский

---

Incomplete human reference genomes can drive false sex biases and expose patient-identifying information in metagenomic data

Nature Communications, Год журнала: 2025, Номер 16(1)

Опубликована: Янв. 18, 2025

Abstract As next-generation sequencing technologies produce deeper genome coverages at lower costs, there is a critical need for reliable computational host DNA removal in metagenomic data. We find that insufficient filtration using prior human references can introduce false sex biases and inadvertently permit flow-through of host-specific during bioinformatic analyses, which could be exploited individual identification. To address these issues, we benchmark three methods varying throughput, with concomitant applications across low biomass samples such as skin high microbial datasets including fecal samples. are important obtaining accurate results (e.g., tissue, skin). Overall, demonstrate rigorous key component privacy-minded analyses patient microbiomes provide computationally efficient pipelines accomplishing this task on large-scale datasets.

Язык: Английский

---

Revisiting the cancer microbiome using PRISM

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2025, Номер unknown

Опубликована: Янв. 24, 2025

Recent controversy around the cancer microbiome highlights need for improved microbial analysis methods human genomics data. We developed PRISM, a computational approach precise microorganism identification and decontamination from low-biomass sequencing PRISM removes spurious signals achieves excellent performance when benchmarked on curated dataset of 62,006 known true- false-positive taxa. then use to detect microbes in 8 types CPTAC TCGA datasets. identify rich microbiomes gastrointestinal tract tumors bacteria subset pancreatic that are associated with altered glycoproteomes, more extensive smoking histories, higher tumor recurrence risk. find relatively sparse other TCGA, which we demonstrate may reflect differing parameters. Overall, does not replace gold-standard controls, but it enables higher-confidence analyses reveals tumor-associated microorganisms potential molecular clinical significance.

Язык: Английский

---

A Systematic Review and Meta-Analysis of 16S rRNA and Cancer Microbiome Atlas Datasets to Characterize Microbiota Signatures in Normal Breast, Mastitis, and Breast Cancer

Microorganisms, Год журнала: 2025, Номер 13(2), С. 467 - 467

Опубликована: Фев. 19, 2025

The breast tissue microbiome has been increasingly recognized as a potential contributor to cancer development and progression. However, inconsistencies in microbial composition across studies have hindered the identification of definitive signatures. We conducted systematic review meta-analysis 11 using 16S rRNA sequencing characterize bacterial 1260 fresh samples, including normal, mastitis-affected, benign, cancer-adjacent, cancerous tissues. Studies published until 31 December 2023 were included if they analyzed human Illumina short-read with sufficient metadata, while non-human non-breast tissues, non-English articles, those lacking metadata or alternative methods excluded. also incorporated data from Cancer Genome Atlas (TCGA-BRCA) cohort enhance our analyses. Our identified Proteobacteria, Firmicutes, Actinobacteriota, Bacteroidota dominant phyla tissue, Staphylococcus Corynebacterium frequently detected studies. While diversity was similar between cancer-adjacent both exhibited lower compared normal mastitis-affected Variability genera observed primer sets studies, emphasizing need for standardized methodologies research. An analysis TCGA-BRCA confirmed dominance Corynebacterium, which associated proliferation-related gene expression programs. Notably, high abundance 4.1-fold increased mortality risk. These findings underscore clinical relevance tumor progression emphasize importance methodological consistency. Future establish causal relationships, elucidate underlying mechanisms, assess microbiome-targeted interventions are warranted.

Язык: Английский

---

Cross-laboratory replication of pseudomyxoma peritonei tumor microbiome reveals reproducible microbial signatures

mSphere, Год журнала: 2025, Номер unknown

Опубликована: Фев. 20, 2025

ABSTRACT Recent work has demonstrated that cancer-specific microbial communities often colonize tumor tissues. However, untangling low-biomass signals from environmental contamination makes this research technically challenging. We utilize pseudomyxoma peritonei (PMP), a cancer characterized by the spread of mucus-secreting cells throughout peritoneal cavity, to develop robust workflow for identifying reproducible microbiomes. Typically originating rupture an appendiceal into metastasized tumors have been previously shown harbor core set microbes. did not control potential these low biomass samples. expand upon prior findings characterizing microbiome 70 additional PMP and six normal tissues along with appropriate laboratory controls. Additionally, DNA subset 25 was extracted sequenced at independent laboratory. found evidence signatures between replicates different include taxa may be introduced surgical contamination, as well patient-specific are also commonly implicated in colorectal cancer. In addition, preoperative chemotherapy treatment reduce diversity. Our demonstrate how sample replication can powerful approach investigate associated will improve research. IMPORTANCE over 30 types The origin their possible involvement carcinogenesis or outcomes remains unclear, yet important area A current major challenge low-biomass, tumor-associated microbiomes is introduction during collection, handling, extraction, PCR, sequencing. Here, we provide framework replicating samples help identify background contamination. Using approach, show (PMP) host communities, including organisms Incorporating future studies promising increase reproducibility field.

Язык: Английский

---

Association of tumor microbiome with survival in resected early-stage PDAC

mSystems, Год журнала: 2025, Номер unknown

Опубликована: Фев. 27, 2025

ABSTRACT The pancreas tumor microbiota may influence microenvironment and survival in early-stage pancreatic ductal adenocarcinoma (PDAC); however, current studies are limited small. We investigated the relationship of to 201 surgically resected patients with localized PDAC (Stages I–II), from Cancer Genome Atlas (TCGA) International Consortium (ICGC) cohorts. characterized microbiome using RNA-sequencing data. examined association overall (OS), via meta-analysis Cox PH model. A microbial risk score (MRS) was calculated OS-associated microbiota. further explored whether is related host immune infiltration. α- β-diversities were not associated OS; 11 bacterial species, including species Gammaproteobacteria , confirmed by extensive resampling, significantly OS (all Q < 0.05). MRS summarizing these bacteria a threefold change (hazard ratio = 2.96 per standard deviation MRS, 95% confidence interval 2.26–3.86). This result consistent across two cohorts stratified analyses adjuvant therapy (chemotherapy/radiation). Identified also exhibited memory B cells naïve CD4 + T cells, which be landscape through BCR TCR signaling pathways. Our study shows that unique structure, potentially affecting microenvironment, poorer PDAC. These findings suggest mechanisms involved survival, informing prognosis guiding personalized treatment strategies. IMPORTANCE Much available data on derived relatively small heterogeneous studies, those involving advanced stages cancer. There critical knowledge gap terms treated surgical resection; we expect advancements initially best achieved who curative intent.

Язык: Английский

---