2021-09-11: ACM Conference on Bioinformatics, Computational Biology, and Health Information (BCB) Trip Report

By -


The ACM Conference on Bioinformatics, Computational Biology, and Health Information (BCB) was held virtually, due to the COVID-19 pandemic, from August 1st through August 4th, 2021.  There were 301 registered attendees this year.  For those that were not lucky enough to attend the conference, the conference proceedings are available online.

This was my first time attending an ACM BCB conference.  It is the flagship conference of the ACM Special Intertest Group on Bioinformatics, Computational Biology, and Biomedical Informatics (SIGBio).  Although I did not get to travel to Chicago where the conference was originally planned to be held, the CONFlux Virtual Platform made it easy to attend the scheduled sessions and interact with other attendees.

There were three general chairs for this year's conference:  Dr. Jiajie Zhang of UTHealth School of Biomedical Informatics, Dr. Hongmei Jiang of Northwestern University, and Dr. Xiuzhen Huang of Arkansas State University.  The conference shared keynote speakers with the 2021 Workshop of Algorithms in Bioinformatics (WABI).  As a result of this and the shared virtual platform, ACM BCB attendees were also able to view presentations that were a part of WABI.

I submitted two late-breaking posters to the conference, which were both accepted:  Visual Pipeline Modeling in NeuroPype for EEG Data Analysis and Sentiment Analysis of Social Media Posts for Mental Health Assessments.  The poster session was on-demand so attendees could view any of the 27 accepted posters at any time during the conference and post questions and/or comments.

Conference Day 0

The first day of the conference fell on a Sunday and was dedicated to workshops and tutorials.  The five workshops centered around a variety of topics to include:  parallel and cloud-based bioinformatics and biomedicine; machine learning models for multi-omics data integration; machine learning methods for single-cell analysis; high-performance computing, big data analytics and integration for multi-omics biomedical data; and non-boundary thinking in bioinformatics.  Tutorials were given on Singularity, Pytorch EHR, computational approaches for 3D genomic modeling, next generation sequencing data analysis, and a graphical front-end for rapid and scalable execution of bioinformatics workflows using serverless cloud computing.  Some sessions were short, lasting only about 50 minutes while one workshop lasted 5.5 hours.  

The day ended with a first-time attendee special session led by Dr. Anna Ritz.  She talked about the ACM BCB conference and how she became involved with it.  She discussed conference etiquette, conference schedules, types of sessions, and settings for the conference virtual platform.  Breakout rooms were created so that conference attendees could meet other conference attendees in smaller groups.  After returning to the main room, Dr. Ritz wrapped up the discussion with additional information about research opportunities, including sharing a site dedicated to computer science research jobs at undergraduate institutions.

Conference Day 1

Morning Sessions

Monday, August 2nd was the first day of the main conference.  It started with three concurrent sessions on Electronic Health Records, Sequence Analysis, and System Biology.

Opening Remarks

After the morning sessions, general and program chairs gave their opening remarks.  Dr. Jiajie Zhang introduced himself and the other two general chairs, Dr. Hongmei Jiang and Dr. Xiuzhen Huang.  He also provided a overview of the conference, identified  conference organizers, and also mentioned sponsors.

Dr. Yu Zhang of Trinity University continued opening remarks with a review of conference statistics. This year, there were 131 submissions from 26 countries with 40 accepted as regular papers and 17 accepted as short papers.  This resulted in an acceptance rate of 30% for regular papers or 43% if also considering short papers.  The conference had 93 program committee members, 104 external reviewers, and 7 expert reviewers.  Dr. Zhang also provided more details on the keynote speakers and featured sessions before giving acknowledgements to all those involved in the conference.

Keynote Speaker 1

The first keynote speaker was Dr. Jason Moore who is a professor of Informatics, Director of the Penn Institute for Biomedical Informatics, and the Senior Associate Dean for Informatics in the Perelman School of Medicine at the University of Pennsylvania.  He has been working for some years with artificial intelligence (AI) and machine learning with application to the biomedical sciences as well as developing automated machine learning packages.  It was, therefore, fitting that Dr. Moore gave a talk on "Accessible Artificial Intelligence for Automating Biomedical Data Science," which included an overview of PennAI.

Afternoon Sessions

The keynote speech was followed by a special session on funding.  Information on different funding sources was presented by Jean Yuan of the National Institutes on Aging, Georgia-Ann Klutke of the National Science Foundation, and Yanli Wang of the National Library of Medicine.

Three concurrent sessions that followed included Genomic Variation, Health Monitoring and Phenotyping, and Structural Bioinformatics.  I chose to attend the session on Health Monitoring and Phenotyping.  The session started with a presentation from Ohio University's Xianlong Zeng on "Transformer-Based Unsupervised Patient Representation Learning Based on Medical Claims for Risk Stratification and Analysis."  That presentation was followed another given by Stony Brook University's Zongxing Xie on "Signal quality detection towards practical non-touch vital sign monitoring." 

Sara Nouri Golmaei from Indiana University-Purdue University continued the session with by sharing work on a hybrid deep learning model as detailed in a paper titled "DeepNote-GNN: Predicting Hospital Readmission using Clinical Notes and Patient Network."

The final presentation of the session came from the University of Utah.  It was given by PhD student Youjia Zhou on "Pheno-Mapper: An Interactive Toolbox For The Visual Exploration Of Phenomics Data."

SIGBio General Meeting

The first day of the conference ended with a SIGBio general meeting that provided attendees with information about SIGBio including membership fees for those who may want to join the group.

Conference Day 2 

Morning Sessions

Day 2 of the main conference started with two sessions:  Single Cell Omics and Machine Learning & Drug Design.  I attended the session on Machine Learning and Drug Design, which consisted of four paper presentations.  However, the virtual platform enabled me to watch some of the recordings from the Single Cell Omics session later.  

In the Machine Learning and Drug Design session, Wake Forest University's Dr. Natalia Khuri presented the first paper on "A value-based approach for training of classifiers with high-throughput small molecule screening data."  Tianfan Fu from the Georgia Institute of Technology presented a paper titled "SPEAR: self-supervised post-training enhancer for molecule optimization."  Simon Fraser University's Amir Hosein Safari also gave a presentation during this session.

Dr. Lizhen Shi from Florida Polytechnic University finished the session with a presentation of her group's work on "LSHvec: a vector representation of DNA sequences using locality sensitive hashing and FastText word embeddings."

Keynote Speaker 2

The second keynote speaker of the conference, Dr. Aidong Zhang, is a William Wulf Faculty Fellow and Professor of Computer Science, with a joint appointment in the Department of Biomedical Engineering and School of Data Science at University of Virginia.  She gave her talk titled "Transfer Learning and Meta Learning for Biomedical Applications" in which described how meta-learning performs better than transfer learning.

Afternoon Sessions

The afternoon started with a featured session on Bioinformatics, Data Science, AI, and COVID.  Dr. Pinar Keskinocak of Georgia Tech started the session with her talk on "Modeling and Analytics for Infectious Diseases and COVID-19."  She talked about prevention, surveillance, and treatment as a part of infectious disease prevention and control as well as the use of disease models for resource estimation and decision making in this process.  Results of Dr. Keskinocak's group's work can be seen on the COVID-19 dashboards used to track status and trends.  Dr. Yuan Luo of Northwestern University followed with a talk on the "Simulation Of Scarce Resource Allocation In Critically Ill Patients With COVID-19."

Dr. Fei Wang gave the final talk of the special session with a presentation on "Data-Driven Subphenotyping For Characterizing The Clinical Heterogeneity Of COVID-19."  His group's aim is to use machine learning to find groups of COVID-19 patients according to their clinical manifestations.  After the three presentations, the panelists responded to questions from the audience. 

The day ended with two concurrent sessions with one centered around Medical Imaging and the other on Graphs and Networks.  I attended the Graphs and Networks session, which consisted of five presentations.  PhD student Jack Lanchantin from UVA started the session with a presentation of his group's work titled "Transfer Learning for Predicting Virus-Host Protein Interactions for Novel Virus Sequences."  Code and data for this work will be available online soon.   Case Western Reserve University's Anuj Godase followed that presentation with his titled "GNNfam: Utilizing Sparsity in Protein Family Prediction using Graph Neural Networks."

Lisa Oh from Dartmouth College discussed her work on "A Multi-Resolution Graph Convolution Network for Contiguous Epitope Prediction."  Anisa Halimi 

Zican Li from the University of Washington Bothell ended the session with a review of his short paper titled "Investigating Statistical Analysis for Network Motifs."

Day 3 

Morning Sessions

The final day of the conference started with two sessions once again.  I decided to attend the live session on COVID-19 and browse recordings of the Clinical Trials and Outcome Prediction session later.  The COVID-19 session consisted of four briefs, with the first one being given by Shayom Debopadhay of the Rensselaer Polytechnic Institute on "Temporal analysis of social determinants associated with COVID-19 mortality."  Tauren Naren of Georgia Tech continued the session with the presentation of her group's paper.

The paper "Surveillance of COVID-19 pandemic using social media: a reddit study in North Carolina" was briefed by Christopher Whitfield of North Carolina A&T University.  Lodewijk Brand presented the last paper of the session titled "A multi-instance support vector machine with incomplete data for clinical outcome prediction of COVID-19."

Keynote Speaker 3

The final keynote speaker of the conference was Dr. Mona Singh from Princeton University's computer science department.  The group she works with focuses on computational molecular biology along with machine learning and algorithms. She gave a presentation on a joint work titled "Algorithms For Deciphering Disease Networks."

Afternoon Sessions

The afternoon began with a featured session on diversity and inclusion: "Challenges and Opportunities."

There was a session held in the afternoon centered around Cancer.  I opted to attend the other session on Ontologies and Databases, which consisted of five presentations.  The work of Lucas Jing Liu and others from the University of Kentucky, "KGDAL: knowledge graph guided double attention LSTM for rolling mortality prediction for AKI-D patients" was the first to be presented.  Dr. Sunil Mohan's group's paper, "Low resource recognition and linking of biomedical concepts from a large ontology" was reviewed before Jiho Noh gave his presentation.

Two more presentations were given before the end of the last regular session of the conference.  Purdue University's Zhuoyan Li presented the paper "HYPON: embedding biomedical ontology with entity sets."  Dr. Ye Wu from the University of Hong Kong discussed his group's work on BioNumQA-BERT.

Awards

Before ending the conference, multiple appreciation awards were given to the conference's technical program chairs, workshop chairs, tutorial chairs, poster chairs, highlights chairs, proceedings chairs, publicity chairs, and its web master.  Acknowledgements were also given to the three key note speakers.

The Best Paper award went to Eric Strobl and Thomas Lasko for their paper titled, "Synthesized Difference in Differences."  There were two best student paper awards.  The first one was for the paper "Match2: Hybrid Self-Organizing Map and Deep Learning Strategies for Treatment Effect Estimation" submitted by Xiao Shou, Tian Gao, Dharmashankar Subramanian, and Kristin Bennett.  Zongxing Xie, Bing Zhou, and Fan Ye also received a Best Student Paper award for their paper "Signal Quality Detection Towards Practical Non-Touch Vital Sign Monitoring."

Best Poster awards were given for three poster submissions:  "Formulating a Gene Signature for Diagnosis of Autoimmune and Infectious Diseases" by Riya Gupta, Aditya Rao, Lara Murphy Jones, Purvesh Khatri; "Mechanistic Model Demonstrates Importance of Autocrine IL-8 Secretion by Neutrophils" by Wangui Mbuguiro, Feilim Mac Gabhann; and "The Genetics of Human Aging: Predicting Age and Age-Related Diseases by Deep Mining High Dimensional Biomarker Data" by Hannah Guan.

Closing Remarks

During closing remarks, conference co-chair Tamer Kahveci informed the audience that the same general chairs would organized the 2022 ACM BCB conference.  Since an in-person conference could not be held this year in Chicago as originally planned, the chairs are planning to hold the 2022 conference there.


-- Bathsheba Farrow (@sheissheba)

Comments

Post a Comment