Off-targetP ML: an open source machine learning framework for off-target panel safety assessment of small molecules (2024)

Naga, D., Muster, W., Musvasva, E. (2022). Off-targetP ML: an open source machine learning framework for off-target panel safety assessment of small molecules. Journal of Cheminformatics, 14(1), [27]. https://doi.org/10.1186/s13321-022-00603-w

Naga, Doha ; Muster, Wolfgang ; Musvasva, Eunice et al. / Off-targetP ML : an open source machine learning framework for off-target panel safety assessment of small molecules. In: Journal of Cheminformatics. 2022 ; Vol. 14, No. 1.

@article{143e8d08284f4f809d066fd99fd73f46,

title = "Off-targetP ML: an open source machine learning framework for off-target panel safety assessment of small molecules",

abstract = "Unpredicted drug safety issues constitute the majority of failures in the pharmaceutical industry according to several studies. Some of these preclinical safety issues could be attributed to the non-selective binding of compounds to targets other than their intended therapeutic target, causing undesired adverse events. Consequently, pharmaceutical companies routinely run in-vitro safety screens to detect off-target activities prior to preclinical and clinical studies. Hereby we present an open source machine learning framework aiming at the prediction of our in-house 50 off-target panel activities for ~ 4000 compounds, directly from their structure. This framework is intended to guide chemists in the drug design process prior to synthesis and to accelerate drug discovery. We also present a set of ML approaches that require minimum programming experience for deployment. The workflow incorporates different ML approaches such as deep learning and automated machine learning. It also accommodates popular issues faced in bioactivity predictions, as data imbalance, inter-target duplicated measurements and duplicated public compound identifiers. Throughout the workflow development, we explore and compare the capability of Neural Networks and AutoML in constructing prediction models for fifty off-targets of different protein classes, different dataset sizes, and high-class imbalance. Outcomes from different methods are compared in terms of efficiency and efficacy. The most important challenges and factors impacting model construction and performance in addition to suggestions on how to overcome such challenges are also discussed.",

keywords = "Drug discovery, Safety screening, Off-target panel, Class imbalance, Deep learning, Automated machine learning (AutoML), Ensembling methods, DRUG ATTRITION, PREDICTION, AUGMENTATION",

author = "Doha Naga and Wolfgang Muster and Eunice Musvasva and Ecker, {Gerhard F.}",

year = "2022",

month = may,

day = "7",

doi = "10.1186/s13321-022-00603-w",

language = "English",

volume = "14",

journal = "Journal of Cheminformatics",

issn = "1758-2946",

publisher = "BioMed Central Ltd.",

number = "1",

}

Naga, D, Muster, W, Musvasva, E 2022, 'Off-targetP ML: an open source machine learning framework for off-target panel safety assessment of small molecules', Journal of Cheminformatics, vol. 14, no. 1, 27. https://doi.org/10.1186/s13321-022-00603-w

Off-targetP ML : an open source machine learning framework for off-target panel safety assessment of small molecules. / Naga, Doha; Muster, Wolfgang; Musvasva, Eunice et al.

In: Journal of Cheminformatics, Vol. 14, No. 1, 27, 07.05.2022.

Publications: Contribution to journal › Article › Peer Reviewed

TY - JOUR

T1 - Off-targetP ML

T2 - an open source machine learning framework for off-target panel safety assessment of small molecules

AU - Naga, Doha

AU - Muster, Wolfgang

AU - Musvasva, Eunice

AU - Ecker, Gerhard F.

PY - 2022/5/7

Y1 - 2022/5/7

N2 - Unpredicted drug safety issues constitute the majority of failures in the pharmaceutical industry according to several studies. Some of these preclinical safety issues could be attributed to the non-selective binding of compounds to targets other than their intended therapeutic target, causing undesired adverse events. Consequently, pharmaceutical companies routinely run in-vitro safety screens to detect off-target activities prior to preclinical and clinical studies. Hereby we present an open source machine learning framework aiming at the prediction of our in-house 50 off-target panel activities for ~ 4000 compounds, directly from their structure. This framework is intended to guide chemists in the drug design process prior to synthesis and to accelerate drug discovery. We also present a set of ML approaches that require minimum programming experience for deployment. The workflow incorporates different ML approaches such as deep learning and automated machine learning. It also accommodates popular issues faced in bioactivity predictions, as data imbalance, inter-target duplicated measurements and duplicated public compound identifiers. Throughout the workflow development, we explore and compare the capability of Neural Networks and AutoML in constructing prediction models for fifty off-targets of different protein classes, different dataset sizes, and high-class imbalance. Outcomes from different methods are compared in terms of efficiency and efficacy. The most important challenges and factors impacting model construction and performance in addition to suggestions on how to overcome such challenges are also discussed.

AB - Unpredicted drug safety issues constitute the majority of failures in the pharmaceutical industry according to several studies. Some of these preclinical safety issues could be attributed to the non-selective binding of compounds to targets other than their intended therapeutic target, causing undesired adverse events. Consequently, pharmaceutical companies routinely run in-vitro safety screens to detect off-target activities prior to preclinical and clinical studies. Hereby we present an open source machine learning framework aiming at the prediction of our in-house 50 off-target panel activities for ~ 4000 compounds, directly from their structure. This framework is intended to guide chemists in the drug design process prior to synthesis and to accelerate drug discovery. We also present a set of ML approaches that require minimum programming experience for deployment. The workflow incorporates different ML approaches such as deep learning and automated machine learning. It also accommodates popular issues faced in bioactivity predictions, as data imbalance, inter-target duplicated measurements and duplicated public compound identifiers. Throughout the workflow development, we explore and compare the capability of Neural Networks and AutoML in constructing prediction models for fifty off-targets of different protein classes, different dataset sizes, and high-class imbalance. Outcomes from different methods are compared in terms of efficiency and efficacy. The most important challenges and factors impacting model construction and performance in addition to suggestions on how to overcome such challenges are also discussed.

KW - Drug discovery

KW - Safety screening

KW - Off-target panel

KW - Class imbalance

KW - Deep learning

KW - Automated machine learning (AutoML)

KW - Ensembling methods

KW - DRUG ATTRITION

KW - PREDICTION

KW - AUGMENTATION

UR - http://www.scopus.com/inward/record.url?scp=85129776890&partnerID=8YFLogxK

U2 - 10.1186/s13321-022-00603-w

DO - 10.1186/s13321-022-00603-w

M3 - Article

VL - 14

JO - Journal of Cheminformatics

JF - Journal of Cheminformatics

SN - 1758-2946

IS - 1

M1 - 27

ER -

Naga D, Muster W, Musvasva E, Ecker GF. Off-targetP ML: an open source machine learning framework for off-target panel safety assessment of small molecules. Journal of Cheminformatics. 2022 May 7;14(1):27. doi: 10.1186/s13321-022-00603-w