Inferring Mechanism of Action of an Unknown Compound from Time Series Omics Data

Vertes, Akos; Arul, Albert-Baskar; Avar, Peter; Korte, Andrew R.; Li, Hang; Nemes, Péter; Parvin, Lida; Stopka, Sylwia A.; Hwang, Sunil; Sahab, Ziad J.; Zhang, Linwen; Bunin, Deborah I.; Knapp, Merrill; Poggio, Andy; Stehr, Mark-Oliver; Talcott, Carolyn; Davis, Brian M.; Dinn, Sean R.; Morton, Christine A.; Sevinsky, Christopher J.; Zavodszky, Maria I.

doi:10.1007/978-3-319-99429-1_14

Cited by 6 publications

(14 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The GP (for Gaussian process) profiles are based on a distribution of continuous log 2 intensity profiles for control and treated samples for each transcript using Gaussian processes (described in [24]). Arrays of means and standard deviations (SD) for control and treated processes were generated by sampling the continuous distributions at 100 time points between 0 and 48 h chosen uniformly on a log scale.…”

Section: Gaussian Process Modelingmentioning

confidence: 99%

“…Clusters were annotated with GO process, GO function, and HUGO Gene Nomenclature Committee (HGNC) family terms using a heuristic method to compute likely classifications of a cluster from classifications of its elements. [24] Ranking. In addition to PCA ranking, autoencoder techniques were used to derive compressed representations of the GP profiles.…”

Section: Gaussian Process Modelingmentioning

confidence: 99%

“…Contrary to the usual criteria, the transcripts that are least accurately recovered are the potentially interesting ones, since they don't behave like most transcripts, i.e. they are anomalies from the autoencoders perspective (see [24]). Generative Adversarial Network (GAN) techniques were used define two additional rankings: atypical/fake and typical/real.…”

Section: Gaussian Process Modelingmentioning

confidence: 99%

“…Regulation intervals delimit times that log 2 fold change stays above 0.75 or below −0.75. The right branch uses time profiles obtained by Gaussian process (GP) modeling [24]. Using these time profiles, transcripts are clustered (k-means using PCA to reduce dimensionality), and ranked by contribution to PCA components and by two machine learning algorithms, one using autoencoder techniques (see [24]) and one using Generative Adversarial Nets (GANs, [6], Section 2.3).…”

Section: Introductionmentioning

confidence: 99%

“…We consider two categories: biological processes, individual transcripts. To identify candidate processes we used PatherDB over representation analysis [13] combined with our GO term annotations of k-means clusters [24]. Potentially key transcripts were selected from the Top20X transcripts using a combination of ranking, clustering, and GO annotations.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Transcriptional Response of SK-N-AS Cells to Methamidophos (Extended Abstract)

Vertes

Arul

Avar

et al. 2019

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Transcriptomics response of SK-N-AS cells to methamidophos (an acetylcholine esterase inhibitor) exposure was measured at 10 time points between 0.5 and 48 h. The data was analyzed using a combination of traditional statistical methods and novel machine learning algorithms for detecting anomalous behavior and infer causal relations between time profiles. We identified several processes that appeared to be upregulated in cells treated with methamidophos including: unfolded protein response, response to cAMP, calcium ion response, and cell-cell signaling. The data confirmed the expected consequence of acetylcholine buildup. In addition, transcripts with potentially key roles were identified and causal networks relating these transcripts were inferred using two different computational methods: Siamese convolutional networks and time warp causal inference. Two types of anomaly detection algorithms, one based on Autoencoders and the other one based on Generative Adversarial Networks (GANs), were applied to narrow down the set of relevant transcripts.

show abstract

Section: Gaussian Process Modelingmentioning

confidence: 99%

Section: Gaussian Process Modelingmentioning

confidence: 99%

Section: Gaussian Process Modelingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations