Log in


  • Home
  • Machine Learning Based Integrative Multi-Omics Analysis— Presented by Global Champion of All Three PrecisionFDA AI Challenges

Machine Learning Based Integrative Multi-Omics Analysis— Presented by Global Champion of All Three PrecisionFDA AI Challenges

  • Fri, February 19, 2021
  • 09:00 - 10:00

Registration


Registration is closed

Fri, Feb 19, 2021, 09:00 – 10:00AM (PDT), 12:00 – 1:00PM (EST)


Abstract:


Sentieon develops highly-optimized algorithms for bioinformatics applications, using the team’s expertise in algorithm, software, and system optimization.  Since 2018, Sentieon participated and won PrecisionFDA’s three multi-omics AI modeling challenges, demonstrating its capability in addressing these emerging problems. In NCI-CPTAC Multi-omics Enabled Sample Mislabeling Correction Challenge (2018), Sentieon’s solution ranked the 1st in both sub-challenges, with perfect score in identifying and correcting all mislabeling. In Brain Cancer Predictive Modeling and Biomarker Discovery Challenge (2019-2020), Sentieon identified the best-performing brain-cancer biomarkers in all three sub-challenges. Moreover, one of Sentieon’s 2020 summer intern projects won precisionFDA’s “VHA Innovation Ecosystem and precisionFDA COVID-19 Risk Factor Modeling Challenge”, in which Sentieon’s three submissions ranked top 3 among all 34 submissions.

Sentieon is applying the machine-learning expertise into multi-omics analysis to enable a broader vision of softPharma. Different from traditional “big-data” machine-learning problems, the key challenge in multi-omics is its unique nature of “small sample with deep data per sample.” 

To address the pain point of time-consuming bioinformatics, Sentieon also developed fast and accurate secondary analysis pipelines that run on generic CPU, covering functions from BCL conversion, alignment, to germline and somatic variant calling. Sentieon DNAseq and TNseq pipelines re-implement BWA/GATK’s mathematics, achieving 10X speedup by designing more efficient computing algorithms. Sentieon also developed DNAscope and TNscope pipelines that feature machine learning to achieve best-in-class variant calling accuracy, and can be easily adapted to many sequencing platforms including Illumina, MGI, and PacBio.

Sentieon pipelines have been adopted by over 300 institutional customers globally, and have more than 100 manuscript citations. Using our expertise in modeling, optimization, machine learning and AI, and high-performance computing, we strive to enable precision data for precision medicine.


Speaker:


Dr. Frank Hu currently serves as Senior Bioinformatics Scientist at Sentieon, responsible for product cycle management and driving Sentieon’s business success in Asian market especially China. Dr. Hu obtained his BS degree from Nanjing University in 2008, PhD in Genomics from The Ohio State University in 2013, and took postdoc training (Computational Biology) at Joint Genome Institute, part of Lawrence Berkeley National Laboratory. He had extensive experience in NGS data analysis, and published multiple manuscripts on peer reviewed journals as first or co-first author. In 2014-2017 Dr. Hu worked at BGI Americas as customer support scientist, moved to Predicine at 2017, and then joined Sentieon in the year of 2018.


Strategic Alliance:


文本转语音功能仅限200个字符
选项 : 历史 : 反馈 : Donate 关闭



@Dahshu 2020

文本转语音功能仅限200个字符
选项 : 历史 : 反馈 : Donate 关闭
G
M
T
Y
文本转语音功能仅限200个字符
Powered by Wild Apricot Membership Software