Medication effects on the gut microbiome in allo-HCT.

Decreased gut microbiome diversity has been associated with negative outcome in allogeneic hematopoietic stem cell transfer (HCT). A study published in this issue of Cell identifies associations between non-antibiotic drug administration, microbiome state transitions, and response to HCT, highlighting the potential impact of such drugs on microbiome and HCT outcome.

The tumour microenvironment shapes innate lymphoid cells in patients with hepatocellular carcinoma.

OBJECTIVE: Hepatocellular carcinoma (HCC) represents a typical inflammation-associated cancer. Tissue resident innate lymphoid cells (ILCs) have been suggested to control tumour surveillance. Here, we studied how the local cytokine milieu controls ILCs in HCC. DESIGN: We performed bulk RNA sequencing of HCC tissue as well as flow cytometry and single-cell RNA sequencing of enriched ILCs from non-tumour liver, margin and tumour core derived from 48 patients with HCC. Simultaneous measurement of protein and RNA expression at the single-cell level (AbSeq) identified precise signatures of ILC subgroups. In vitro culturing of ILCs was used to validate findings from in silico analysis. Analysis of RNA-sequencing data from large HCC cohorts allowed stratification and survival analysis based on transcriptomic signatures. RESULTS: RNA sequencing of tumour, non-tumour and margin identified tumour-dependent gradients, which were associated with poor survival and control of ILC plasticity. Single-cell RNA sequencing and flow cytometry of ILCs from HCC livers identified natural killer (NK)-like cells in the non-tumour tissue, losing their cytotoxic profile as they transitioned into tumour ILC1 and NK-like-ILC3 cells. Tumour ILC composition was mediated by cytokine gradients that directed ILC plasticity towards activated tumour ILC2s. This was liver-specific and not seen in ILCs from peripheral blood mononuclear cells. Patients with high ILC2/ILC1 ratio expressed interleukin-33 in the tumour that promoted ILC2 generation, which was associated with better survival. CONCLUSION: Our results suggest that the tumour cytokine milieu controls ILC composition and HCC outcome. Specific changes of cytokines modify ILC composition in the tumour by inducing plasticity and alter ILC function.

Tumor heterogeneity assessed by sequencing and fluorescence in situ hybridization (FISH) data.

MOTIVATION: Computational reconstruction of clonal evolution in cancers has become a crucial tool for understanding how tumors initiate and progress and how this process varies across patients. The field still struggles, however, with special challenges of applying phylogenetic methods to cancers, such as the prevalence and importance of copy number alteration (CNA) and structural variation events in tumor evolution, which are difficult to profile accurately by prevailing sequencing methods in such a way that subsequent reconstruction by phylogenetic inference algorithms is accurate. RESULTS: In this work, we develop computational methods to combine sequencing with multiplex interphase fluorescence in situ hybridization to exploit the complementary advantages of each technology in inferring accurate models of clonal CNA evolution accounting for both focal changes and aneuploidy at whole-genome scales. By integrating such information in an integer linear programming framework, we demonstrate on simulated data that incorporation of FISH data substantially improves accurate inference of focal CNA and ploidy changes in clonal evolution from deconvolving bulk sequence data. Analysis of real glioblastoma data for which FISH, bulk sequence and single cell sequence are all available confirms the power of FISH to enhance accurate reconstruction of clonal copy number evolution in conjunction with bulk and optionally single-cell sequence data. AVAILABILITY AND IMPLEMENTATION: Source code is available on Github at https://github.com/CMUSchwartzLab/FISH_deconvolution. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Integration of adeno-associated virus (AAV) into the genomes of most Thai and Mongolian liver cancer patients does not induce oncogenesis.

BACKGROUND: Engineered versions of adeno-associated virus (AAV) are commonly used in gene therapy but evidence revealing a potential oncogenic role of natural AAV in hepatocellular carcinoma (HCC) has raised concerns. The frequency of potentially oncogenic integrations has been reported in only a few populations. AAV infection and host genome integration in another type of liver cancer, cholangiocarcinoma (CCA), has been studied only in one cohort. All reported oncogenic AAV integrations in HCC come from strains resembling the fully sequenced AAV2 and partly sequenced AAV13. When AAV integration occurs, only a fragment of the AAV genome is detectable in later DNA or RNA sequencing. The integrated fragment is typically from the 3' end of the AAV genome, and this positional bias has been only partly explained. Three research groups searched for evidence of AAV integration in HCC RNAseq samples in the Cancer Genome Atlas (TCGA) but reported conflicting results. RESULTS: We collected and analyzed whole transcriptome and viral capture DNA sequencing in paired tumor and non-tumor samples from two liver cancer Asian cohorts from Thailand (N = 147, 47 HCC and 100 intrahepatic cholangiocarcinoma (iCCA)) and Mongolia (N = 70, all HCC). We found only one HCC patient with a potentially oncogenic integration of AAV, in contrast to higher frequency reported in European patients. There were no oncogenic AAV integrations in iCCA patients. AAV genomic segments are present preferentially in the non-tumor samples of Thai patients. By analyzing the AAV genome positions of oncogenic and non-oncogenic integrated fragments, we found that almost all the putative oncogenic integrations overlap the X gene, which is present and functional only in the strain AAV2 among all fully sequenced strains. This gene content difference could explain why putative oncogenic integrations from other AAV strains have not been reported. We resolved the discrepancies in previous analyses of AAV presence in TCGA HCC samples and extended it to CCA. There are 12 TCGA samples with an AAV segment and none are in Asian patients. AAV segments are present in preferentially in TCGA non-tumor samples, like what we observed in the Thai patients. CONCLUSIONS: Our findings suggest a minimal AAV risk of hepatocarcinogenesis in Asian liver cancer patients. The partial genome presence and positional bias of AAV integrations into the human genome has complicated analysis of possible roles of AAV in liver cancer.

PhISCS-BnB: a fast branch and bound algorithm for the perfect tumor phylogeny reconstruction problem.

MOTIVATION: Recent advances in single-cell sequencing (SCS) offer an unprecedented insight into tumor emergence and evolution. Principled approaches to tumor phylogeny reconstruction via SCS data are typically based on general computational methods for solving an integer linear program, or a constraint satisfaction program, which, although guaranteeing convergence to the most likely solution, are very slow. Others based on Monte Carlo Markov Chain or alternative heuristics not only offer no such guarantee, but also are not faster in practice. As a result, novel methods that can scale up to handle the size and noise characteristics of emerging SCS data are highly desirable to fully utilize this technology. RESULTS: We introduce PhISCS-BnB (phylogeny inference using SCS via branch and bound), a branch and bound algorithm to compute the most likely perfect phylogeny on an input genotype matrix extracted from an SCS dataset. PhISCS-BnB not only offers an optimality guarantee, but is also 10-100 times faster than the best available methods on simulated tumor SCS data. We also applied PhISCS-BnB on a recently published large melanoma dataset derived from the sublineages of a cell line involving 20 clones with 2367 mutations, which returned the optimal tumor phylogeny in <4 h. The resulting phylogeny agrees with and extends the published results by providing a more detailed picture on the clonal evolution of the tumor. AVAILABILITY AND IMPLEMENTATION: https://github.com/algo-cancer/PhISCS-BnB. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

High Levels of Chromosomal Copy Number Alterations and TP53 Mutations Correlate with Poor Outcome in Younger Breast Cancer Patients.

Prognosis in young patients with breast cancer is generally poor, yet considerable differences in clinical outcomes between individual patients exist. To understand the genetic basis of the disparate clinical courses, tumors were collected from 34 younger women, 17 with good and 17 with poor outcomes, as determined by disease-specific survival during a follow-up period of 17 years. The clinicopathologic parameters of the tumors were complemented with DNA image cytometry profiles, enumeration of copy numbers of eight breast cancer genes by multicolor fluorescence in situ hybridization, and targeted sequence analysis of 563 cancer genes. Both groups included diploid and aneuploid tumors. The degree of intratumor heterogeneity was significantly higher in aneuploid versus diploid cases, and so were gains of the oncogenes MYC and ZNF217. Significantly more copy number alterations were observed in the group with poor outcome. Almost all tumors in the group with long survival were classified as luminal A, whereas triple-negative tumors predominantly occurred in the short survival group. Mutations in PIK3CA were more common in the group with good outcome, whereas TP53 mutations were more frequent in patients with poor outcomes. This study shows that TP53 mutations and the extent of genomic imbalances are associated with poor outcome in younger breast cancer patients and thus emphasize the central role of genomic instability vis-a-vis tumor aggressiveness.

Aneuploidy, TP53 mutation, and amplification of MYC correlate with increased intratumor heterogeneity and poor prognosis of breast cancer patients.

The clinical course of breast cancer varies from one patient to another. Currently, the choice of therapy relies on clinical parameters and histological and molecular tumor features. Alas, these markers are informative in only a subset of patients. Therefore, additional predictors of disease outcome would be valuable for treatment stratification. Extensive studies showed that the degree of variation of the nuclear DNA content, i.e., aneuploidy, determines prognosis. Our aim was to further elucidate the molecular basis of aneuploidy. We analyzed five diploid and six aneuploid tumors with more than 20 years of follow-up. By performing FISH with a multiplexed panel of 10 probes to enumerate copy numbers in individual cells, and by sequencing 563 cancer-related genes, we analyzed how aneuploidy is linked to intratumor heterogeneity. In our cohort, none of the patients with diploid tumors died of breast cancer during follow-up in contrast to four of six patients with aneuploid tumors (mean survival 86.4 months). The FISH analysis showed markedly increased genomic instability and intratumor heterogeneity in aneuploid tumors. MYC gain was observed in only 20% of the diploid cancers, while all aneuploid cases showed a gain. The mutation burden was similar in diploid and aneuploid tumors, however, TP53 mutations were not observed in diploid tumors, but in all aneuploid tumors in our collective. We conclude that quantitative measurements of intratumor heterogeneity by multiplex FISH, detection of MYC amplification and TP53 mutation could augment prognostication in breast cancer patients.

A common SLC26A4-linked haplotype underlying non-syndromic hearing loss with enlargement of the vestibular aqueduct.

BACKGROUND: Enlargement of the vestibular aqueduct (EVA) is the most common radiological abnormality in children with sensorineural hearing loss. Mutations in coding regions and splice sites of the SLC26A4 gene are often detected in Caucasians with EVA. Approximately one-fourth of patients with EVA have two mutant alleles (M2), one-fourth have one mutant allele (M1) and one-half have no mutant alleles (M0). The M2 genotype is correlated with a more severe phenotype. METHODS: We performed genotype-haplotype analysis and massively parallel sequencing of the SLC26A4 region in patients with M1 EVA and their families. RESULTS: We identified a shared novel haplotype, termed CEVA (Caucasian EVA), composed of 12 uncommon variants upstream of SLC26A4. The presence of the CEVA haplotype on seven of ten 'mutation-negative' chromosomes in a National Institutes of Health M1 EVA discovery cohort and six of six mutation-negative chromosomes in a Danish M1 EVA replication cohort is higher than the observed prevalence of 28 of 1006 Caucasian control chromosomes (p<0.0001 for each EVA cohort). The corresponding heterozygous carrier rate is 28/503 (5.6%). The prevalence of CEVA (11 of 126) is also increased among M0 EVA chromosomes (p=0.0042). CONCLUSIONS: The CEVA haplotype causally contributes to most cases of Caucasian M1 EVA and, possibly, some cases of M0 EVA. The CEVA haplotype of SLC26A4 defines the most common allele associated with hereditary hearing loss in Caucasians. The diagnostic yield and prognostic utility of sequence analysis of SLC26A4 exons and splice sites will be markedly increased by addition of testing for the CEVA haplotype.

Phylogenetic analysis of multiple FISH markers in oral tongue squamous cell carcinoma suggests that a diverse distribution of copy number changes is associated with poor prognosis.

Oral tongue squamous cell carcinoma (OTSCC) is associated with poor prognosis. To improve prognostication, we analyzed four gene probes (TERC, CCND1, EGFR and TP53) and the centromere probe CEP4 as a marker of chromosomal instability, using fluorescence in situ hybridization (FISH) in single cells from the tumors of sixty-five OTSCC patients (Stage I, n = 15; Stage II, n = 30; Stage III, n = 7; Stage IV, n = 13). Unsupervised hierarchical clustering of the FISH data distinguished three clusters related to smoking status. Copy number increases of all five markers were found to be correlated to non-smoking habits, while smokers in this cohort had low-level copy number gains. Using the phylogenetic modeling software FISHtrees, we constructed models of tumor progression for each patient based on the four gene probes. Then, we derived test statistics on the models that are significant predictors of disease-free and overall survival, independent of tumor stage and smoking status in multivariate analysis. The patients whose tumors were modeled as progressing by a more diverse distribution of copy number changes across the four genes have poorer prognosis. This is consistent with the view that multiple genetic pathways need to become deregulated in order for cancer to progress.

Inferring models of multiscale copy number evolution for single-tumor phylogenetics.

MOTIVATION: Phylogenetic algorithms have begun to see widespread use in cancer research to reconstruct processes of evolution in tumor progression. Developing reliable phylogenies for tumor data requires quantitative models of cancer evolution that include the unusual genetic mechanisms by which tumors evolve, such as chromosome abnormalities, and allow for heterogeneity between tumor types and individual patients. Previous work on inferring phylogenies of single tumors by copy number evolution assumed models of uniform rates of genomic gain and loss across different genomic sites and scales, a substantial oversimplification necessitated by a lack of algorithms and quantitative parameters for fitting to more realistic tumor evolution models. RESULTS: We propose a framework for inferring models of tumor progression from single-cell gene copy number data, including variable rates for different gain and loss events. We propose a new algorithm for identification of most parsimonious combinations of single gene and single chromosome events. We extend it via dynamic programming to include genome duplications. We implement an expectation maximization (EM)-like method to estimate mutation-specific and tumor-specific event rates concurrently with tree reconstruction. Application of our algorithms to real cervical cancer data identifies key genomic events in disease progression consistent with prior literature. Classification experiments on cervical and tongue cancer datasets lead to improved prediction accuracy for the metastasis of primary cervical cancers and for tongue cancer survival. AVAILABILITY AND IMPLEMENTATION: Our software (FISHtrees) and two datasets are available at ftp://ftp.ncbi.nlm.nih.gov/pub/FISHtrees.

Inherited biallelic CSF3R mutations in severe congenital neutropenia.

Severe congenital neutropenia (SCN) is characterized by low numbers of peripheral neutrophil granulocytes and a predisposition to life-threatening bacterial infections. We describe a novel genetic SCN type in 2 unrelated families associated with recessively inherited loss-of-function mutations in CSF3R, encoding the granulocyte colony-stimulating factor (G-CSF) receptor. Family A, with 3 affected children, carried a homozygous missense mutation (NM_000760.3:c.922C>T, NP_000751.1:p.Arg308Cys), which resulted in perturbed N-glycosylation and aberrant localization to the cell surface. Family B, with 1 affected infant, carried compound heterozygous deletions provoking frameshifts and premature stop codons (NM_000760.3:c.948_963del, NP_000751.1:p.Gly316fsTer322 and NM_000760.3:c.1245del, NP_000751.1:p.Gly415fsTer432). Despite peripheral SCN, all patients had morphologic evidence of full myeloid cell maturation in bone marrow. None of the patients responded to treatment with recombinant human G-CSF. Our study highlights the genetic and morphologic SCN variability and provides evidence both for functional importance and redundancy of G-CSF receptor-mediated signaling in human granulopoiesis.

Potential non-B DNA regions in the human genome are associated with higher rates of nucleotide mutation and expression variation.

While individual non-B DNA structures have been shown to impact gene expression, their broad regulatory role remains elusive. We utilized genomic variants and expression quantitative trait loci (eQTL) data to analyze genome-wide variation propensities of potential non-B DNA regions and their relation to gene expression. Independent of genomic location, these regions were enriched in nucleotide variants. Our results are consistent with previously observed mutagenic properties of these regions and counter a previous study concluding that G-quadruplex regions have a reduced frequency of variants. While such mutagenicity might undermine functionality of these elements, we identified in potential non-B DNA regions a signature of negative selection. Yet, we found a depletion of eQTL-associated variants in potential non-B DNA regions, opposite to what might be expected from their proposed regulatory role. However, we also observed that genes downstream of potential non-B DNA regions showed higher expression variation between individuals. This coupling between mutagenicity and tolerance for expression variability of downstream genes may be a result of evolutionary adaptation, which allows reconciling mutagenicity of non-B DNA structures with their location in functionally important regions and their potential regulatory role.

Deleterious mutations in LRBA are associated with a syndrome of immune deficiency and autoimmunity.

Most autosomal genetic causes of childhood-onset hypogammaglobulinemia are currently not well understood. Most affected individuals are simplex cases, but both autosomal-dominant and autosomal-recessive inheritance have been described. We performed genetic linkage analysis in consanguineous families affected by hypogammaglobulinemia. Four consanguineous families with childhood-onset humoral immune deficiency and features of autoimmunity shared genotype evidence for a linkage interval on chromosome 4q. Sequencing of positional candidate genes revealed that in each family, affected individuals had a distinct homozygous mutation in LRBA (lipopolysaccharide responsive beige-like anchor protein). All LRBA mutations segregated with the disease because homozygous individuals showed hypogammaglobulinemia and autoimmunity, whereas heterozygous individuals were healthy. These mutations were absent in healthy controls. Individuals with homozygous LRBA mutations had no LRBA, had disturbed B cell development, defective in vitro B cell activation, plasmablast formation, and immunoglobulin secretion, and had low proliferative responses. We conclude that mutations in LRBA cause an immune deficiency characterized by defects in B cell activation and autophagy and by susceptibility to apoptosis, all of which are associated with a clinical phenotype of hypogammaglobulinemia and autoimmunity.

The phenotype of human STK4 deficiency.

We describe a novel clinical phenotype associating T- and B-cell lymphopenia, intermittent neutropenia, and atrial septal defects in 3 members of a consanguineous kindred. Their clinical histories included recurrent bacterial infections, viral infections, mucocutaneous candidiasis, cutaneous warts, and skin abscesses. Homozygosity mapping and candidate gene sequencing revealed a homozygous premature termination mutation in the gene STK4 (serine threonine kinase 4, formerly having the symbol MST1). STK4 is the human ortholog of Drosophila Hippo, the central constituent of a highly conserved pathway controlling cell growth and apoptosis. STK4-deficient lymphocytes and neutrophils exhibit enhanced loss of mitochondrial membrane potential and increased susceptibility to apoptosis. STK4 deficiency is a novel human primary immunodeficiency syndrome.

Chromosome 7 to the rescue: overcoming chromosome 10 loss in gliomas.

The co-occurrence of chromosome 10 loss and chromosome 7 gain in gliomas is the most frequent loss-gain co-aneuploidy pair in human cancers, a phenomenon that has been investigated without resolution since the late 1980s. Expanding beyond previous gene-centric studies, we investigate the co-occurrence in a genome-wide manner taking an evolutionary perspective. First, by mining large tumor aneuploidy data, we predict that the more likely order is 10 loss followed by 7 gain. Second, by analyzing extensive genomic and transcriptomic data from both patients and cell lines, we find that this co-occurrence can be explained by functional rescue interactions that are highly enriched on 7, which can possibly compensate for any detrimental consequences arising from the loss of 10. Finally, by analyzing transcriptomic data from normal, non-cancerous, human brain tissues, we provide a plausible reason why this co-occurrence happens preferentially in cancers originating in certain regions of the brain.

Identification of intracellular bacteria from multiple single-cell RNA-seq platforms using CSI-Microbes.

The study of the tumor microbiome has been garnering increased attention. We developed a computational pipeline (CSI-Microbes) for identifying microbial reads from single-cell RNA sequencing (scRNA-seq) data and for analyzing differential abundance of taxa. Using a series of controlled experiments and analyses, we performed the first systematic evaluation of the efficacy of recovering microbial unique molecular identifiers by multiple scRNA-seq technologies, which identified the newer 10x chemistries (3' v3 and 5') as the best suited approach. We analyzed patient esophageal and colorectal carcinomas and found that reads from distinct genera tend to co-occur in the same host cells, testifying to possible intracellular polymicrobial interactions. Microbial reads are disproportionately abundant within myeloid cells that up-regulate proinflammatory cytokines like IL1Β and CXCL8, while infected tumor cells up-regulate antigen processing and presentation pathways. These results show that myeloid cells with bacteria engulfed are a major source of bacterial RNA within the tumor microenvironment (TME) and may inflame the TME and influence immunotherapy response.

Linkage analysis of a large African family segregating stuttering suggests polygenic inheritance.

We describe a pedigree of 71 individuals from the Republic of Cameroon in which at least 33 individuals have a clinical diagnosis of persistent stuttering. The high concentration of stuttering individuals suggests that the pedigree either contains a single highly penetrant gene variant or that assortative mating led to multiple stuttering-associated variants being transmitted in different parts of the pedigree. No single locus displayed significant linkage to stuttering in initial genome-wide scans with microsatellite and SNP markers. By dividing the pedigree into five subpedigrees, we found evidence for linkage to previously reported loci on 3q and 15q, and to novel loci on 2p, 3p, 14q, and a different region of 15q. Using the two-locus mode of Superlink, we showed that combining the recessive locus on 2p and a single-locus additive representation of the 15q loci is sufficient to achieve a two-locus score over 6 on the entire pedigree. For this 2p + 15q analysis, we show LOD scores ranging from 4.69 to 6.57, and the scores are sensitive to which marker is chosen for 15q. Our findings provide strong evidence for linkage at several loci.

Accuracy and coverage assessment of Oryctolagus cuniculus (rabbit) genes encoding immunoglobulins in the whole genome sequence assembly (OryCun2.0) and localization of the IGH locus to chromosome 20.

We report on the analyses of genes encoding immunoglobulin heavy and light chains in the rabbit 6.51× whole genome assembly. This OryCun2.0 assembly confirms previous mapping of the duplicated IGK1 and IGK2 loci to chromosome 2 and the IGL lambda light chain locus to chromosome 21. The most frequently rearranged and expressed IGHV1 that is closest to IG DH and IGHJ genes encodes rabbit VHa allotypes. The partially inbred Thorbecke strain rabbit used for whole-genome sequencing was homozygous at the IGK but heterozygous with the IGHV1a1 allele in one of 79 IGHV-containing unplaced scaffolds and IGHV1a2, IGHM, IGHG, and IGHE sequences in another. Some IGKV, IGLV, and IGHA genes are also in other unplaced scaffolds. By fluorescence in situ hybridization, we assigned the previously unmapped IGH locus to the q-telomeric region of rabbit chromosome 20. An approximately 3-Mb segment of human chromosome 14 including IGH genes predicted to map to this telomeric region based on synteny analysis could not be located on assembled chromosome 20. Unplaced scaffold chrUn0053 contains some of the genes that comparative mapping predicts to be missing. We identified discrepancies between previous targeted studies and the OryCun2.0 assembly and some new BAC clones with IGH sequences that can guide other studies to further sequence and improve the OryCun2.0 assembly. Complete knowledge of gene sequences encoding variable regions of rabbit heavy, kappa, and lambda chains will lead to better understanding of how and why rabbits produce antibodies of high specificity and affinity through gene conversion and somatic hypermutation.

Inflammatory bowel disease and mutations affecting the interleukin-10 receptor.

BACKGROUND: The molecular cause of inflammatory bowel disease is largely unknown. METHODS: We performed genetic-linkage analysis and candidate-gene sequencing on samples from two unrelated consanguineous families with children who were affected by early-onset inflammatory bowel disease. We screened six additional patients with early-onset colitis for mutations in two candidate genes and carried out functional assays in patients' peripheral-blood mononuclear cells. We performed an allogeneic hematopoietic stem-cell transplantation in one patient. RESULTS: In four of nine patients with early-onset colitis, we identified three distinct homozygous mutations in genes IL10RA and IL10RB, encoding the IL10R1 and IL10R2 proteins, respectively, which form a heterotetramer to make up the interleukin-10 receptor. The mutations abrogate interleukin-10-induced signaling, as shown by deficient STAT3 (signal transducer and activator of transcription 3) phosphorylation on stimulation with interleukin-10. Consistent with this observation was the increased secretion of tumor necrosis factor alpha and other proinflammatory cytokines from peripheral-blood mononuclear cells from patients who were deficient in IL10R subunit proteins, suggesting that interleukin-10-dependent "negative feedback" regulation is disrupted in these cells. The allogeneic stem-cell transplantation performed in one patient was successful. CONCLUSIONS: Mutations in genes encoding the IL10R subunit proteins were found in patients with early-onset enterocolitis, involving hyperinflammatory immune responses in the intestine. Allogeneic stem-cell transplantation resulted in disease remission in one patient.

Predicting cancer prognosis and drug response from the tumor microbiome.

Tumor gene expression is predictive of patient prognosis in some cancers. However, RNA-seq and whole genome sequencing data contain not only reads from host tumor and normal tissue, but also reads from the tumor microbiome, which can be used to infer the microbial abundances in each tumor. Here, we show that tumor microbial abundances, alone or in combination with tumor gene expression, can predict cancer prognosis and drug response to some extent-microbial abundances are significantly less predictive of prognosis than gene expression, although similarly as predictive of drug response, but in mostly different cancer-drug combinations. Thus, it appears possible to leverage existing sequencing technology, or develop new protocols, to obtain more non-redundant information about prognosis and drug response from RNA-seq and whole genome sequencing experiments than could be obtained from tumor gene expression or genomic data alone.