Jie Fan

Jie Fan


Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
Verified
Journal Article
Open access
PDF file
KLKe Liu
Ke Liu
XSXiangyan Sun
Xiangyan Sun
LJLei Jia
Lei Jia
JMJun Ma
Jun Ma
HXHaoming Xing
Haoming Xing
JFJie Fan
Jie Fan
Absorption, distribution, metabolism, and excretion (ADME) studies are critical for drug discovery. Conventionally, these tasks, together with other chemical property predictions, rely on domain-specific feature descriptors, or fingerprints. Following the recent success of neural networks, we developed Chemi-Net, a completely data-driven, domain knowledge-free, deep learning method for ADME property prediction. To compare the relative performance of Chemi-Net with Cubist, one of the popular machine learning programs used by Amgen, a large-scale ADME property prediction study was performed on-site at Amgen. For all 13 data sets, Chemi-Net resulted in higher R2 values compared with the Cubist benchmark. The median R2 increase rate over Cubist was 26.7%. We expect that the significantly increased accuracy of ADME prediction seen with Chemi-Net over Cubist will greatly accelerate drug discovery.

17 December 2025 in Nature Communications
Verified
Journal Article
Open access
PDF file
EHErika Hamilton
Erika Hamilton
RLRachel M Layman
Rachel M Layman
DCDavid Cosgrove
David Cosgrove
MDMichael Danso
Michael Danso
HZHui Zhang
Hui Zhang
MPManish R. Patel
Manish R. Patel
AC699 is a novel, orally bioavailable chimeric estrogen receptor-α (ERα) degrader that induces proteasome-dependent ERα degradation via cereblon E3 ligase recruitment. Here we report findings from a first-in-human, phase 1, dose-escalation study of once-daily AC699 (100–600 mg) in patients with heavily-pretreated, locally-advanced/metastatic ER-positive/HER2-negative breast cancer (ClinicalTrials.gov Identifier: NCT05654532). Primary objectives are to assess dose-limiting toxicities and treatment-emergent adverse events (TEAEs). Secondary objectives are to evaluate pharmacokinetics, objective response rate (ORR), clinical benefit rate (CBR, including stable disease ≥24 weeks), duration of response (DOR), and progression-free survival (PFS). Among 37 treated patients, TEAEs occurred in 78% of patients, most commonly nausea (19%), fatigue (16%), and neutropenia (16%). All treatment-related adverse events were Grade 1/2, with no dose reductions/discontinuations; the maximum tolerated dose was not reached. Of 26 efficacy-evaluable patients, 4 (15%) achieved partial responses and CBR was 23%. In exploratory analysis of patients with ESR1 mutations, ORR and CBR were 40% and 45%, respectively. Median DOR and PFS were 6.5 and 3.6 months overall, and 6.5 and 7.4 months in ESR1-mutant patients. AC699 steady-state exposure increased approximately dose-proportionally between 100–400 mg, plateauing at 600 mg. AC699 demonstrated favorable safety, predictable PK, and encouraging antitumor activity, particularly in ESR1-mutant disease.

21 March 2018 in arXiv
Verified
Preprint
Open access
PDF file
KLKe Liu
Ke Liu
XSXiangyan Sun
Xiangyan Sun
LJLei Jia
Lei Jia
JMJun Ma
Jun Ma
HXHaoming Xing
Haoming Xing
JFJie Fan
Jie Fan
Absorption, distribution, metabolism, and excretion (ADME) studies are critical for drug discovery. Conventionally, these tasks, together with other chemical property predictions, rely on domain-specific feature descriptors, or fingerprints. Following the recent success of neural networks, we developed Chemi-Net, a completely data-driven, domain knowledge-free, deep learning method for ADME property prediction. To compare the relative performance of Chemi-Net with Cubist, one of the popular machine learning programs used by Amgen, a large-scale ADME property prediction study was performed on-site at Amgen. The results showed that our deep neural network method improved current methods by a large margin. We foresee that the significantly increased accuracy of ADME prediction seen with Chemi-Net over Cubist will greatly accelerate drug discovery.

19 April 2021 in arXiv
Verified
Preprint
Open access
PDF file
KLKe Liu
Ke Liu
ZNZekun Ni
Zekun Ni
ZZZhenyu Zhou
Zhenyu Zhou
STSuocheng Tan
Suocheng Tan
XZXun Zou
Xun Zou
JFJie Fan
Jie Fan
Molecular modeling is an important topic in drug discovery. Decades of research have led to the development of high quality scalable molecular force fields. In this paper, we show that neural networks can be used to train a universal approximator for energy potential functions. By incorporating a fully automated training process we have been able to train smooth, differentiable, and predictive potential functions on large-scale crystal structures. A variety of tests have also been performed to show the superiority and versatility of the machine-learned model.

26 July 2017 in arXiv
Verified
Preprint
Open access
PDF file
KLKe Liu
Ke Liu
XSXiangyan Sun
Xiangyan Sun
JMJun Ma
Jun Ma
ZZZhenyu Zhou
Zhenyu Zhou
QDQilin Dong
Qilin Dong
JFJie Fan
Jie Fan
A deep neural network based architecture was constructed to predict amino acid side chain conformation with unprecedented accuracy. Amino acid side chain conformation prediction is essential for protein homology modeling and protein design. Current widely-adopted methods use physics-based energy functions to evaluate side chain conformation. Here, using a deep neural network architecture without physics-based assumptions, we have demonstrated that side chain conformation prediction accuracy can be improved by more than 25%, especially for aromatic residues compared with current standard methods. More strikingly, the prediction method presented here is robust enough to identify individual conformational outliers from high resolution structures in a protein data bank without providing its structural factors. We envisage that our amino acid side chain predictor could be used as a quality check step for future protein structure model validation and many other potential applications such as side chain assignment in Cryo-electron microscopy, crystallography model auto-building, protein folding and small molecule ligand docking.

11 September 2025 in Research Square
Verified
Preprint
Open access
PDF file
JFJie Fan
Jie Fan
EHErika Hamilton
Erika Hamilton
RLRachel M Layman
Rachel M Layman
DCDavid Cosgrove
David Cosgrove
MDMichael Danso
Michael Danso
MPManish Patel
Manish Patel
AC699, an oral, novel chimeric estrogen receptor-α (ERα) binds, ubiquitinates, and degrades ERα by engaging cereblon E3 ligase. We report AC699 (100‒600 mg daily), safety/tolerability, and preliminary antitumor efficacy from a first-in-human, phase 1, dose-escalation study in 37 patients with heavily-pretreated, locally-advanced/metastatic ER+/HER2- breast cancer. 78% of patients had treatment-emergent adverse events (TEAEs), most commonly nausea (19%), fatigue (16%), and neutrophil count decreased (16%). All treatment-related AEs (TRAE) were Grade 1/2. No TRAE-associated dose reductions/discontinuations occurred. The maximum tolerated dose was not reached. Among 26 evaluable patients, 4 (15%) achieved confirmed partial response (PR) and the clinical benefit rate (CBR; including stable disease ≥24 weeks) was 23%. Corresponding values in patients with ESR1 mutations were 40% (4/10 patients, all PRs) and 45% (5/11 patients). These results demonstrate promising safety/tolerability, and antitumor activity in patients with ER+/HER2‒ advanced/metastatic breast cancer, with notable PRs in 40% of patients with ESR1 mutations.