Dataset Information

Latent-space embedding of expression data identifies gene signatures from sputum samples of asthmatic patients.

ABSTRACT:

Background

The pathogenesis of asthma is a complex process involving multiple genes and pathways. Identifying biomarkers from asthma datasets, especially those that include heterogeneous subpopulations, is challenging. Potentially, autoencoders provide ideal frameworks for such tasks as they can embed complex, noisy high-dimensional gene expression data into a low-dimensional latent space in an unsupervised fashion, enabling us to extract distinguishing features from expression data.

Results

Here, we developed a framework combining a denoising autoencoder and a supervised learning classifier to identify gene signatures related to asthma severity. Using the trained autoencoder with 50 hidden units, we found that hierarchical clustering on the low-dimensional embedding corresponds well with previously defined and clinically relevant clusters of patients. Moreover, each hidden unit has contributions from each of the genes, and pathway analysis of these contributions shows that the hidden units are significantly enriched in known asthma-related pathways. We then used genes that contribute most to the hidden units to develop a secondary random-forest classifier for directly predicting asthma severity. The feature importance metric from this classifier identified a signature based on 50 key genes, which are associated with severity. Furthermore, we can use these key genes to successfully estimate FEV1/FVC ratios across patients, via support-vector-machine regression.

Conclusion

We found that the denoising autoencoder framework can extract meaningful patterns corresponding to functional gene groups and patient clusters from the gene expression of asthma patients.

SUBMITTER: Lou S

PROVIDER: S-EPMC7560063 | biostudies-literature | 2020 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Latent-space embedding of expression data identifies gene signatures from sputum samples of asthmatic patients.

Lou Shaoke S Li Tianxiao T Spakowicz Daniel D Yan Xiting X Chupp Geoffrey Lowell GL Gerstein Mark M

BMC bioinformatics 20201015 1

<h4>Background</h4>The pathogenesis of asthma is a complex process involving multiple genes and pathways. Identifying biomarkers from asthma datasets, especially those that include heterogeneous subpopulations, is challenging. Potentially, autoencoders provide ideal frameworks for such tasks as they can embed complex, noisy high-dimensional gene expression data into a low-dimensional latent space in an unsupervised fashion, enabling us to extract distinguishing features from expression data.<h4> ...[more]

PMID: 33059594

Similar Datasets

Project description:BackgroundChronic obstructive pulmonary disease (COPD) is a lung disease characterised by airflow-limiting inflammation and mucus production. Acute exacerbations are a major cause of COPD-related morbidity and mortality and are mostly associated with bacterial or viral infections. A vaccine targeting non-typeable Haemophilus influenzae (NTHi) and Moraxella catarrhalis (Mcat), the main bacteria associated with exacerbations, was tested in a Phase 2 trial. We assessed "ex-vivo" expression of vaccine candidate and housekeeping genes pd, pe, pilA, gapA, ompP6 of NTHi, and uspA2, parE, polA of Mcat in sputum samples of COPD patients and determined whether expression of the vaccine candidate genes pd, pe, pilA (NTHi) and uspA2 (Mcat) differed between stable and exacerbation samples.MethodsA single-centre, prospective, observational cohort study was conducted where 123 COPD patients were seen on enrolment, followed monthly for 2 years, and reviewed after onset of acute exacerbations. We selected 69 patients with sputum samples positive for NTHi or Mcat by PCR during at least one stable and one exacerbation visit. mRNA was isolated from the sputum, and expression of NTHi and Mcat genes was analysed with RT-PCR. Statistical analyses compared mRNA concentrations between stable and exacerbation samples and in relationship to COPD severity and exacerbation frequency.ResultsThe vaccine candidate genes were variably expressed in sputum samples, suggesting they are expressed in the lung. Absolute and relative expression of all NTHi vaccine candidate genes and Mcat uspA2 were similar between exacerbation and stable samples. Expression of pd and pilA was slightly associated with the number of exacerbations in the year before enrolment, and uspA2 with the disease severity status at enrolment.ConclusionsThe NTHi-Mcat vaccine candidate genes were expressed in sputum samples, and each gene had a specific level of expression. No statistically significant differences in gene expression were detectable between stable and exacerbation samples. However, the history of COPD exacerbations was slightly associated with the expression of pd, pilA and uspA2. Trial registration NCT01360398 ( https://www.Clinicaltrialsgov ).

Dataset Information

Latent-space embedding of expression data identifies gene signatures from sputum samples of asthmatic patients.

Background

Results

Conclusion

Publications

Latent-space embedding of expression data identifies gene signatures from sputum samples of asthmatic patients.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets