ICON: An Adaptation of Infinite HMMs for Time Traces with Drift

Ioannis Sgouralis, Steve Presse

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

Bayesian nonparametric methods have recently transformed emerging areas within data science. One such promising method, the infinite hidden Markov model (iHMM), generalizes the HMM that itself has become a workhorse in single molecule data analysis. The iHMM goes beyond the HMM by self-consistently learning all parameters learned by the HMM in addition to learning the number of states without recourse to any model selection steps. Despite its generality, simple features (such as drift), common to single molecule time traces, result in an overinterpretation of drift and the introduction of artifact states. Here we present an adaptation of the iHMM that can treat data with drift originating from one or many traces (e.g., Förster resonance energy transfer). Our fully Bayesian method couples the iHMM to a continuous control process (drift) self-consistently learned while learning all other quantities determined by the iHMM (including state numbers). A key advantage of this method is that all traces—regardless of drift or states visited across traces—may now be treated on an equal footing, thereby eliminating user-dependent trace selection (based on drift levels), preprocessing to remove drift, and postprocessing model selection based on state number.

Original languageEnglish (US)
Pages (from-to)2117-2126
Number of pages10
JournalBiophysical Journal
Volume112
Issue number10
DOIs
StatePublished - May 23 2017

Fingerprint

Bayes Theorem
Learning
Energy Transfer
Artifacts

ASJC Scopus subject areas

  • Biophysics

Cite this

ICON : An Adaptation of Infinite HMMs for Time Traces with Drift. / Sgouralis, Ioannis; Presse, Steve.

In: Biophysical Journal, Vol. 112, No. 10, 23.05.2017, p. 2117-2126.

Research output: Contribution to journalArticle

Sgouralis, Ioannis ; Presse, Steve. / ICON : An Adaptation of Infinite HMMs for Time Traces with Drift. In: Biophysical Journal. 2017 ; Vol. 112, No. 10. pp. 2117-2126.
@article{113a99800a2145dbb8e2a8246fb88d1c,
title = "ICON: An Adaptation of Infinite HMMs for Time Traces with Drift",
abstract = "Bayesian nonparametric methods have recently transformed emerging areas within data science. One such promising method, the infinite hidden Markov model (iHMM), generalizes the HMM that itself has become a workhorse in single molecule data analysis. The iHMM goes beyond the HMM by self-consistently learning all parameters learned by the HMM in addition to learning the number of states without recourse to any model selection steps. Despite its generality, simple features (such as drift), common to single molecule time traces, result in an overinterpretation of drift and the introduction of artifact states. Here we present an adaptation of the iHMM that can treat data with drift originating from one or many traces (e.g., F{\"o}rster resonance energy transfer). Our fully Bayesian method couples the iHMM to a continuous control process (drift) self-consistently learned while learning all other quantities determined by the iHMM (including state numbers). A key advantage of this method is that all traces—regardless of drift or states visited across traces—may now be treated on an equal footing, thereby eliminating user-dependent trace selection (based on drift levels), preprocessing to remove drift, and postprocessing model selection based on state number.",
author = "Ioannis Sgouralis and Steve Presse",
year = "2017",
month = "5",
day = "23",
doi = "10.1016/j.bpj.2017.04.009",
language = "English (US)",
volume = "112",
pages = "2117--2126",
journal = "Biophysical Journal",
issn = "0006-3495",
publisher = "Biophysical Society",
number = "10",

}

TY - JOUR

T1 - ICON

T2 - An Adaptation of Infinite HMMs for Time Traces with Drift

AU - Sgouralis, Ioannis

AU - Presse, Steve

PY - 2017/5/23

Y1 - 2017/5/23

N2 - Bayesian nonparametric methods have recently transformed emerging areas within data science. One such promising method, the infinite hidden Markov model (iHMM), generalizes the HMM that itself has become a workhorse in single molecule data analysis. The iHMM goes beyond the HMM by self-consistently learning all parameters learned by the HMM in addition to learning the number of states without recourse to any model selection steps. Despite its generality, simple features (such as drift), common to single molecule time traces, result in an overinterpretation of drift and the introduction of artifact states. Here we present an adaptation of the iHMM that can treat data with drift originating from one or many traces (e.g., Förster resonance energy transfer). Our fully Bayesian method couples the iHMM to a continuous control process (drift) self-consistently learned while learning all other quantities determined by the iHMM (including state numbers). A key advantage of this method is that all traces—regardless of drift or states visited across traces—may now be treated on an equal footing, thereby eliminating user-dependent trace selection (based on drift levels), preprocessing to remove drift, and postprocessing model selection based on state number.

AB - Bayesian nonparametric methods have recently transformed emerging areas within data science. One such promising method, the infinite hidden Markov model (iHMM), generalizes the HMM that itself has become a workhorse in single molecule data analysis. The iHMM goes beyond the HMM by self-consistently learning all parameters learned by the HMM in addition to learning the number of states without recourse to any model selection steps. Despite its generality, simple features (such as drift), common to single molecule time traces, result in an overinterpretation of drift and the introduction of artifact states. Here we present an adaptation of the iHMM that can treat data with drift originating from one or many traces (e.g., Förster resonance energy transfer). Our fully Bayesian method couples the iHMM to a continuous control process (drift) self-consistently learned while learning all other quantities determined by the iHMM (including state numbers). A key advantage of this method is that all traces—regardless of drift or states visited across traces—may now be treated on an equal footing, thereby eliminating user-dependent trace selection (based on drift levels), preprocessing to remove drift, and postprocessing model selection based on state number.

UR - http://www.scopus.com/inward/record.url?scp=85019904674&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85019904674&partnerID=8YFLogxK

U2 - 10.1016/j.bpj.2017.04.009

DO - 10.1016/j.bpj.2017.04.009

M3 - Article

C2 - 28538149

AN - SCOPUS:85019904674

VL - 112

SP - 2117

EP - 2126

JO - Biophysical Journal

JF - Biophysical Journal

SN - 0006-3495

IS - 10

ER -