Phenotypic Screening Combined with Machine Learning for Efficient Identification of Breast Cancer-Selective Therapeutic Targets

Prson Gautam, Alok Jaiswal, Tero Aittokallio, Hassan Al-Ali, Krister Wennerberg

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

The lack of functional understanding of most mutations in cancer, combined with the non-druggability of most proteins, challenge genomics-based identification of oncology drug targets. We implemented a machine-learning-based approach (idTRAX), which relates cell-based screening of small-molecule compounds to their kinase inhibition data, to directly identify effective and readily druggable targets. We applied idTRAX to triple-negative breast cancer cell lines and efficiently identified cancer-selective targets. For example, we found that inhibiting AKT selectively kills MFM-223 and CAL148 cells, while inhibiting FGFR2 only kills MFM-223. Since the effects of catalytically inhibiting a protein can diverge from those of reducing its levels, targets identified by idTRAX frequently differ from those identified through gene knockout/knockdown methods. This is critical if the purpose is to identify targets specifically for small-molecule drug development, whereby idTRAX may produce fewer false-positives. The rapid nature of the approach suggests that it may be applicable in personalizing therapy.

Original languageEnglish (US)
Pages (from-to)970-979.e4
JournalCell Chemical Biology
Volume26
Issue number7
DOIs
StatePublished - Jul 18 2019

Fingerprint

Learning systems
Identification (control systems)
Screening
Triple Negative Breast Neoplasms
Breast Neoplasms
Gene Knockdown Techniques
Gene Knockout Techniques
Molecules
Oncology
Genomics
Pharmaceutical Preparations
Neoplasms
Proteins
Phosphotransferases
Genes
Cells
Cell Line
Mutation
Therapeutics
Machine Learning

Keywords

  • AI
  • Akt
  • AURKA
  • cancer cell line
  • dependency
  • drug screening
  • FGFR
  • gene silencing
  • kinase
  • kinase inhibitors
  • machine learning
  • PKIS
  • target deconvolution
  • TNBC

ASJC Scopus subject areas

  • Biochemistry
  • Molecular Medicine
  • Molecular Biology
  • Pharmacology
  • Drug Discovery
  • Clinical Biochemistry

Cite this

Phenotypic Screening Combined with Machine Learning for Efficient Identification of Breast Cancer-Selective Therapeutic Targets. / Gautam, Prson; Jaiswal, Alok; Aittokallio, Tero; Al-Ali, Hassan; Wennerberg, Krister.

In: Cell Chemical Biology, Vol. 26, No. 7, 18.07.2019, p. 970-979.e4.

Research output: Contribution to journalArticle

Gautam, Prson ; Jaiswal, Alok ; Aittokallio, Tero ; Al-Ali, Hassan ; Wennerberg, Krister. / Phenotypic Screening Combined with Machine Learning for Efficient Identification of Breast Cancer-Selective Therapeutic Targets. In: Cell Chemical Biology. 2019 ; Vol. 26, No. 7. pp. 970-979.e4.
@article{07f889e47be649f0a67e9ada2ef462da,
title = "Phenotypic Screening Combined with Machine Learning for Efficient Identification of Breast Cancer-Selective Therapeutic Targets",
abstract = "The lack of functional understanding of most mutations in cancer, combined with the non-druggability of most proteins, challenge genomics-based identification of oncology drug targets. We implemented a machine-learning-based approach (idTRAX), which relates cell-based screening of small-molecule compounds to their kinase inhibition data, to directly identify effective and readily druggable targets. We applied idTRAX to triple-negative breast cancer cell lines and efficiently identified cancer-selective targets. For example, we found that inhibiting AKT selectively kills MFM-223 and CAL148 cells, while inhibiting FGFR2 only kills MFM-223. Since the effects of catalytically inhibiting a protein can diverge from those of reducing its levels, targets identified by idTRAX frequently differ from those identified through gene knockout/knockdown methods. This is critical if the purpose is to identify targets specifically for small-molecule drug development, whereby idTRAX may produce fewer false-positives. The rapid nature of the approach suggests that it may be applicable in personalizing therapy.",
keywords = "AI, Akt, AURKA, cancer cell line, dependency, drug screening, FGFR, gene silencing, kinase, kinase inhibitors, machine learning, PKIS, target deconvolution, TNBC",
author = "Prson Gautam and Alok Jaiswal and Tero Aittokallio and Hassan Al-Ali and Krister Wennerberg",
year = "2019",
month = "7",
day = "18",
doi = "10.1016/j.chembiol.2019.03.011",
language = "English (US)",
volume = "26",
pages = "970--979.e4",
journal = "Cell Chemical Biology",
issn = "2451-9448",
publisher = "Elsevier Inc.",
number = "7",

}

TY - JOUR

T1 - Phenotypic Screening Combined with Machine Learning for Efficient Identification of Breast Cancer-Selective Therapeutic Targets

AU - Gautam, Prson

AU - Jaiswal, Alok

AU - Aittokallio, Tero

AU - Al-Ali, Hassan

AU - Wennerberg, Krister

PY - 2019/7/18

Y1 - 2019/7/18

N2 - The lack of functional understanding of most mutations in cancer, combined with the non-druggability of most proteins, challenge genomics-based identification of oncology drug targets. We implemented a machine-learning-based approach (idTRAX), which relates cell-based screening of small-molecule compounds to their kinase inhibition data, to directly identify effective and readily druggable targets. We applied idTRAX to triple-negative breast cancer cell lines and efficiently identified cancer-selective targets. For example, we found that inhibiting AKT selectively kills MFM-223 and CAL148 cells, while inhibiting FGFR2 only kills MFM-223. Since the effects of catalytically inhibiting a protein can diverge from those of reducing its levels, targets identified by idTRAX frequently differ from those identified through gene knockout/knockdown methods. This is critical if the purpose is to identify targets specifically for small-molecule drug development, whereby idTRAX may produce fewer false-positives. The rapid nature of the approach suggests that it may be applicable in personalizing therapy.

AB - The lack of functional understanding of most mutations in cancer, combined with the non-druggability of most proteins, challenge genomics-based identification of oncology drug targets. We implemented a machine-learning-based approach (idTRAX), which relates cell-based screening of small-molecule compounds to their kinase inhibition data, to directly identify effective and readily druggable targets. We applied idTRAX to triple-negative breast cancer cell lines and efficiently identified cancer-selective targets. For example, we found that inhibiting AKT selectively kills MFM-223 and CAL148 cells, while inhibiting FGFR2 only kills MFM-223. Since the effects of catalytically inhibiting a protein can diverge from those of reducing its levels, targets identified by idTRAX frequently differ from those identified through gene knockout/knockdown methods. This is critical if the purpose is to identify targets specifically for small-molecule drug development, whereby idTRAX may produce fewer false-positives. The rapid nature of the approach suggests that it may be applicable in personalizing therapy.

KW - AI

KW - Akt

KW - AURKA

KW - cancer cell line

KW - dependency

KW - drug screening

KW - FGFR

KW - gene silencing

KW - kinase

KW - kinase inhibitors

KW - machine learning

KW - PKIS

KW - target deconvolution

KW - TNBC

UR - http://www.scopus.com/inward/record.url?scp=85068833898&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068833898&partnerID=8YFLogxK

U2 - 10.1016/j.chembiol.2019.03.011

DO - 10.1016/j.chembiol.2019.03.011

M3 - Article

C2 - 31056464

AN - SCOPUS:85068833898

VL - 26

SP - 970-979.e4

JO - Cell Chemical Biology

JF - Cell Chemical Biology

SN - 2451-9448

IS - 7

ER -