Edward W. Huang

Edward W. Huang

Auto-generated profile, based on publicly available information

Edward W. Huang

Amazon, Palo Alto, CA, USA

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

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

Representative Publications

View All Publications
22 January 2020 in PLOS Computational Biology
Verified
Journal Article
Open access
PDF file
EHEdward W Huang
Edward W Huang
ABAmeya Bhope
Ameya Bhope
JLJing Lim
Jing Lim
SSSaurabh Sinha
Saurabh Sinha
AEAmin Emad
Amin Emad
Prediction of clinical drug response (CDR) of cancer patients, based on their clinical and molecular profiles obtained prior to administration of the drug, can play a significant role in individualized medicine. Machine learning models have the potential to address this issue but training them requires data from a large number of patients treated with each drug, limiting their feasibility. While large databases of drug response and molecular profiles of preclinical in-vitro cancer cell lines (CCLs) exist for many drugs, it is unclear whether preclinical samples can be used to predict CDR of real patients. We designed a systematic approach to evaluate how well different algorithms, trained on gene expression and drug response of CCLs, can predict CDR of patients. Using data from two large databases, we evaluated various linear and non-linear algorithms, some of which utilized information on gene interactions. Then, we developed a new algorithm called TG-LASSO that explicitly integrates information on samples’ tissue of origin with gene expression profiles to improve prediction performance. Our results showed that regularized regression methods provide better prediction performance. However, including the network information or common methods of including information on the tissue of origin did not improve the results. On the other hand, TG-LASSO improved the predictions and distinguished resistant and sensitive patients for 7 out of 13 drugs. Additionally, TG-LASSO identified genes associated with the drug response, including known targets and pathways involved in the drugs’ mechanism of action. Moreover, genes identified by TG-LASSO for multiple drugs in a tissue were associated with patient survival. In summary, our analysis suggests that preclinical samples can be used to predict CDR of patients and identify biomarkers of drug sensitivity and survival. Cancer is among the leading causes of death globally and prediction of the drug response of patients to different treatments based on their clinical and molecular profiles can enable individualized cancer medicine. Machine learning algorithms have the potential to play a significant role in this task; however, these algorithms are designed based on the premise that a large number of labeled training samples are available, and these samples are accurate representations of the profiles of real tumors. However, due to ethical and technical reasons, it is not possible to screen humans for many drugs, significantly limiting the size of training data. To overcome this data scarcity problem, machine learning models can be trained using large databases of preclinical samples (e.g. cancer cell line cultures). However, due to the major differences between preclinical samples and real tumors, it is unclear how accurately such preclinical-to-clinical computational models can predict the clinical drug response of cancer patients. Here, first we systematically evaluate a variety of different linear and nonlinear machine learning algorithms for this particular task using two large databases of preclinical (GDSC) and tumor samples (TCGA). Then, we present a novel method called TG-LASSO that utilizes a new approach for explicitly incorporating the tissue of origin of samples in the prediction task. Our results show that TG-LASSO outperforms all other algorithms and can distinguish resistant and sensitive patients for the majority of the tested drugs. Follow-up analysis reveal that this method can also identify biomarkers of drug sensitivity in each cancer type.

13 May 2024
Verified
Conference Paper
Open access
CDCharles Dickens
Charles Dickens
EHEdward Huang
Edward Huang
ARAishwarya Reganti
Aishwarya Reganti
JZJiong Zhu
Jiong Zhu
KSKarthik Subbian
Karthik Subbian
DKDanai Koutra
Danai Koutra
Graph summarization as a preprocessing step is an effective and complementary technique for scalable graph neural network (GNN) training. In this work, we propose the Coarsening Via Convolution Matching (ConvMatch) algorithm and a highly scalable variant, A-ConvMatch, for creating summarized graphs that preserve the output of graph convolution. We evaluate ConvMatch on six real-world link prediction and node classification graph datasets, and show it is efficient and preserves prediction performance while significantly reducing the graph size. Notably, ConvMatch achieves up to 95% of the prediction performance of GNNs on node classification while trained on graphs summarized down to 1% the size of the original graph. Furthermore, on link prediction tasks, ConvMatch consistently outperforms all baselines, achieving up to a 2X improvement.

20 March 2019 in PLOS Computational Biology
Verified
Journal Article
Open access
PDF file
SWSheng Wang
Sheng Wang
EHEdward Huang
Edward Huang
JCJunmei Cairns
Junmei Cairns
JPJian Peng
Jian Peng
LWLiewei Wang
Liewei Wang
SSSaurabh Sinha
Saurabh Sinha
Basal gene expression levels have been shown to be predictive of cellular response to cytotoxic treatments. However, such analyses do not fully reveal complex genotype- phenotype relationships, which are partly encoded in highly interconnected molecular networks. Biological pathways provide a complementary way of understanding drug response variation among individuals. In this study, we integrate chemosensitivity data from a large-scale pharmacogenomics study with basal gene expression data from the CCLE project and prior knowledge of molecular networks to identify specific pathways mediating chemical response. We first develop a computational method called PACER, which ranks pathways for enrichment in a given set of genes using a novel network embedding method. It examines a molecular network that encodes known gene-gene as well as gene-pathway relationships, and determines a vector representation of each gene and pathway in the same low-dimensional vector space. The relevance of a pathway to the given gene set is then captured by the similarity between the pathway vector and gene vectors. To apply this approach to chemosensitivity data, we identify genes whose basal expression levels in a panel of cell lines are correlated with cytotoxic response to a compound, and then rank pathways for relevance to these response-correlated genes using PACER. Extensive evaluation of this approach on benchmarks constructed from databases of compound target genes and large collections of drug response signatures demonstrates its advantages in identifying compound-pathway associations compared to existing statistical methods of pathway enrichment analysis. The associations identified by PACER can serve as testable hypotheses on chemosensitivity pathways and help further study the mechanisms of action of specific cytotoxic drugs. More broadly, PACER represents a novel technique of identifying enriched properties of any gene set of interest while also taking into account networks of known gene-gene relationships and interactions. Gene expression levels have been used to study the cellular response to drug treatments. However, analysis of gene expression without considering gene interactions cannot fully reveal complex genotype-phenotype relationships. Biological pathways reveal the interactions among genes, thus providing a complementary way of understanding the drug response variation among individuals. In this paper, we aim to identify pathways that mediate the chemical response of each drug. We used the recently generated CTRP pharmacogenomics data and CCLE basal expression data to identify these pathways. We showed that using the prior knowledge encoded in molecular networks substantially improves pathway identification. In particular, we integrate genes and pathways into a large heterogeneous network in which links are protein-protein interactions and gene-pathway affiliations. We then project this heterogeneous network onto a low-dimensional space, which enables more precise similarity measurements between pathways and drug-response-correlated genes. Extensive experiments on two benchmarks show that our method substantially improved the pathway identification performance by using the molecular networks. More importantly, our method represents a novel technique of identifying enriched properties of any gene set of interest while also taking into account networks of known gene-gene relationships and interactions.

27 February 2023
Verified
Conference Paper
Open access
YHYan Han
Yan Han
EHEdward W. Huang
Edward W. Huang
WZWenqing Zheng
Wenqing Zheng
NRNikhil Rao
Nikhil Rao
ZWZhangyang (Atlas) Wang
Zhangyang (Atlas) Wang
KSKarthik Subbian
Karthik Subbian
At E-Commerce stores such as Amazon, eBay, and Taobao, the shopping items and the query words that customers use to search for the items form a bipartite graph that captures search behavior. Such a query-item graph can be used to forecast search trends or improve search results. For example, generating query-item associations, which is equivalent to predicting links in the bipartite graph, can yield recommendations that can customize and improve the user search experience. Although the bipartite shopping graphs are straightforward to model search behavior, they suffer from two challenges: 1) The majority of items are sporadically searched and hence have noisy/sparse query associations, leading to a long-tail distribution. 2) Infrequent queries are more likely to link to popular items, leading to another hurdle known as disassortative mixing. To address these two challenges, we go beyond the bipartite graph to take a hypergraph perspective, introducing a new paradigm that leverages auxiliary information from anonymized customer engagement sessions to assist the main task of query-item link prediction. This auxiliary information is available at web scale in the form of search logs. We treat all items appearing in the same customer session as a single hyperedge. The hypothesis is that items in a customer session are unified by a common shopping interest. With these hyperedges, we augment the original bipartite graph into a new hypergraph. We develop a Dual-Channel Attention-Based Hypergraph Neural Network (DCAH), which synergizes information from two potentially noisy sources (original query-item edges and item-item hyperedges). In this way, items on the tail are better connected due to the extra hyperedges, thereby enhancing their link prediction performance. We further integrate DCAH with self-supervised graph pre-training and/or DropEdge training, both of which effectively alleviate disassortative mixing. Extensive experiments on three proprietary E-Commerce datasets show that DCAH yields significant improvements of up to 24.6% in mean reciprocal rank (MRR) and 48.3% in recall compared to GNN-based baselines. Our source code is available at https://github.com/amazon-science/dual-channel-hypergraph-neural-network.

28 December 2024 in ACM Transactions on Knowledge Discovery From Data
Verified
Journal Article
Open access
JZJiong Zhu
Jiong Zhu
ARAishwarya Reganti
Aishwarya Reganti
EHEdward W. Huang
Edward W. Huang
CDCharles Dickens
Charles Dickens
NRNikhil Rao
Nikhil Rao
DKDanai Koutra
Danai Koutra
Distributed graph neural network (GNN) training facilitates learning on massive graphs that surpass the storage and computational capabilities of a single machine. Traditional distributed frameworks strive for performance parity with centralized training by maximally recovering cross-instance node dependencies, relying either on inter-instance communication or periodic fallback to centralized training. However, these processes create overhead and constrain the scalability of the framework. In this work, we propose a streamlined framework for distributed GNN training that eliminates these costly operations, yielding improved scalability, convergence speed, and performance over state-of-the-art approaches. Our framework (1) comprises independent trainers that asynchronously learn local models from locally available parts of the training graph and (2) synchronizes these local models only through periodic (time-based) model aggregation. Contrary to prevailing belief, our theoretical analysis shows that it is not essential to maximize the recovery of cross-instance node dependencies to achieve performance parity with centralized training. Instead, our framework leverages randomized assignment of nodes or super-nodes (i.e., collections of original nodes) to partition the training graph in order to enhance data uniformity and minimize discrepancies in gradient and loss function across instances. Experiments on social and e-commerce networks with up to 1.3 billion edges show that our proposed framework achieves state-of-the-art performance and 2.31 \(\times\) speedup compared to the fastest baseline despite using less training data.

2 October 2016
Verified
Conference Paper
Open access
EHEdward W. Huang
Edward W. Huang
Sheng WangSheng Wang
Sheng Wang
RZRunshun Zhang
Runshun Zhang
BLBaoyan Liu
Baoyan Liu
XZXuezhong Zhou
Xuezhong Zhou
CZChengXiang Zhai
ChengXiang Zhai
Traditional Chinese medicine (TCM), a style of medicine widely used in China for thousands of years, can complement modern western medicine by taking personalization as the core principle of clinical practice. A fundamental task in TCM, particularly important for achieving effective precision medicine, is to subcategorize patients with a general disease into groups corresponding to variations of that disease. In this paper, we conduct the first study of the problem of subcategorizing electronic patient records in TCM. While the general problem of subcategorization can be solved using basic clustering algorithms, accommodating variations in symptoms and herb prescriptions of TCM patient records when computing patient similarity is a major technical challenge that has yet to be addressed. To tackle this problem, we propose to learn inexact matchings of both symptoms and herbs from a TCM dictionary of herb functions by using an embedding algorithm. Our hypothesis is that the prior knowledge of herb-symptom associations in the TCM dictionary can be used to discover latent relationships among comorbid symptoms and functionally similar herbs, thereby improving the quality of subcategorization. We performed extensive experiments on large-scale real-world datasets. As expected, our approach leads to more accurate matchings between patient records than baseline approaches, and thus better subcategorization results. We also show that the proposed algorithm can be used immediately in multiple clinical applications, such as retrieving similar patients as well as discovering two special TCM cases: similar symptoms treated by different herbs and different symptoms treated by similar herbs.