Sign In

Serafim Batzoglou

Genomics, Proteomics, ML

Serafim Batzoglou is the Chief Data Officer at Seer Inc.12 Prior to this, he served as Chief Data Officer at insitro, VP of computational genomics at Illumina, and a professor of computer science at Stanford University from 2001 to 2016.2 His work focuses on computational genomics, with emphasis on creating algorithms, machine learning methods, and systems for analyzing large-scale genomic data. Batzoglou has also contributed to the Human Genome Project and ENCODE.2

Background and Education:: Batzoglou completed his undergraduate studies at MIT and earned a Ph.D. in Computer Science from MIT in 2000, under the supervision of Bonnie Berger.256 He also credits Sorin Istrail and David McAllester as early mentors.6

Career Highlights::

  • Seer Inc.: Current Chief Data Officer (October 2022–Present)1
  • insitro: Former Chief Data Officer13
  • Illumina: Former VP of Applied and Computational Biology13
  • Stanford University: Professor of Computer Science from 2001 to 201623
  • DNAnexus: Co-founder of DNAnexus, a secure cloud platform for scientific collaboration.13
  • Early Career: Lead algorithms designer and implementer of ARACHNE, one of the first programs for whole genome shotgun sequence assembly.56

Achievements and Awards::

  • ISCB Fellow (2020)2
  • ISCB Innovator Award (2016)25
  • Sloan Research Fellowship, Alfred P. Sloan Foundation2
  • Career Award in Computer Science, National Science Foundation2
  • Top 100 Young Technology Innovators, MIT Technology Review23
  • Best Paper Award, ISMB (2003)2

Batzoglou's research interests include alignment algorithms, comparative genomics, gene regulation, regulatory motif finding, and microarray analysis.4 He has also been involved in developing genome alignment tools such as LAGAN.56

Highlights

Feb 2 · linkedin

Our latest paper on pQTLs using Seer technology, led by Karsten Suhre, just went online in Nature Comms: https://lnkd.in/ecuiRXaD

Protein quantitiative trait loci (pQTLs) are genomic variants that affect the level of protein abundance, for example in blood. Performing pQTL genome-wide discovery has important applications, including in understand the direction of causation in linking genes to phenotypes: is a the activity of a gene causal to a phenotype, or a downstream symptom of the phenotype?

One confounder in pQTL analysis is variants within proteins, which are correlated to nearby genome variants (i.e., they are in linkage disaquilibrium) while simultaneously, potentially, affecting the measured abundance of the protein due to the technology used for measurement. For example, if an antibody is used to fish out the protein, and the antibody binds in the location of the variant, the presence of the variant will affect the measured (but not necessarily the true) protein abundance.

Mass spectrometry allows overcoming this difficulty, but only if the computational analysis properly accounts for protein variants. Suhre and colleagues (and I enjoyed being one of the colleagues) developed the computational pipeline to do so, and here show an initial demo of pQTL analysis in the Qatar Diabetes cohort.

Related Questions

What are Serafim Batzoglou's main responsibilities as Chief Data Officer at Seer?
How did Serafim Batzoglou's work at Stanford University influence his career?
What are some notable projects Serafim Batzoglou has led at Illumina?
How did Serafim Batzoglou's role at DNAnexus evolve over time?
What awards and recognitions has Serafim Batzoglou received throughout his career?
Serafim Batzoglou
Serafim Batzoglou, photo 1
Serafim Batzoglou, photo 2
Add to my network

Location

Palo Alto, California, United States