[don.git] / resume / research_statement.mdwn

# Research Objectives

## Uncovering genetic causes of diseases using a multidisciplinary bioinformatics-driven approach



My research focuses on designing and using bioinformatics techniques
to identify causes and mechanisms underlying human diseases such as
Systemic Lupus Erythematosus, Glioblastoma Multiforme, and
Arteriovenous Malformations followed be designing appropriate
diagnostics and treatment methods. Once identified, I work with
collaborators to verify the bioinformatics-discovered mechanisms
utilizing *in silico*, *in vitro*, and *in vivo* techniques, and
develop therapeutic and diagnostic techniques to identify and treat
the underlying human disorder. Cell-culture-based methods are utilized
as the first step in designing multi-target treatments, followed by
appropriate animal models..

# Unique Qualifications

## Straddling Biology, Computer Science, and Statistics

In addition to being a cellular and molecular biologist, I have
extensive experience in algorithm design, computer programming, and
statistics. The combination of these areas enables me to handle
biological problems which involve large numbers of samples and data
and require statistical analysis which can address confounders which
are often present in non-laboratory settings. It also gives me a
unique perspective which enables me to design novel methods to analyze
and interpret large amounts of data.

## Multidisciplinary approach

I have published in multiple disciplines, including membrane
biophysics, bioinformatics, genetics, and cellular biology. My
experiences in these fields and my experiences while transitioning
fields has allowed me to apply unique insights garnered from my
previous work to new topics leading to novel approaches and
breakthrough discoveries.

# Previous Research Projects



## Genetic Basis of System Lupus Erythematosus (SLE)

I developed novel bioinformatic methods which increase 


likelihood of identifying reproducible genetic associations using
prior knowledge from publicly available databases and expert
information [#Armstrong2008:function2gene]. Using these methods, I was
able to identify genes previously unassociated with SLE in a
trio-based study [#Jacob2007:ar_lupus]. These genes were then
replicated in a larger case-control study which was funded by the NIH
on the basis of the original findings [#Jacob2009:sle_irak1] ,
[#Armstrong2009:sle_gi]. Among other findings, this larger study
identified a missense allele in NCF2 (H389Q, rs17849502) which was
associated with SLE. Collaborative work indicated that H389Q altered
the binding energy of NCF2 with VAV1 using docking simulations, and in
vitro experiments confirmed that H389Q altered NADPH oxidase function
[#Jacob2012:sle_ncf2], thus identifying it as a causative SLE
mutation.

## Cancer Stem Cells in Glioblastoma

Glioblastoma is a almost invariably fatal form of brain cancer[^survival_rate] which is
diagnosed in \(\approx\) 9,000 people in the US annually. It is typified
by high levels of chemotherapeutic-resistant recurrences, some of
which is likely caused by cancer stem cells which are insensitive to
many chemotherapeutic agents. In collaboration with Florence Hoffman
at USC, I have classified glioblastoma-derived cancer stem cells and
cancer cell lines into distinct classes using gene microarrays and
various clustering approaches, which will enable the design of a
class-specific treatment of this devastating disease.

## Regulatory pathways underlying cranial arteriovenous malformations

The mechanisms underlying the formation of arteriovenous
malformations[^avm_definition] (AVMs) which occur in the brain are
unknown. Using gene microarrays, qtPCR, and *in vitro* experiments on
primary human endothelial cells cultured from resected AVMs, I
indentified multiple gene regulation pathways, many of them novel,
including the Id1/Thsb1 inhibitory pathway. Additional experiments
indicated that the *in vitro* pathology of endothelial cells could be
partially rescued using extracellular Thsb1
[#Stapleton2011:thbs1].

# Future research directions

## Identification of causal mutations in SLE

While many regions and SNPs which are associated with SLE have been
identified, few of those regions have identified causal alleles with
known function. Furthermore, even for regions with identified causal
alleles, no systematic searches have been performed to identify
additional causal regions. Continuing my existing collaboration with
Chaim O. Jacob, I will rectify this by

1.  Deep sequencing the associated regions in 500 lupus cases and 1000 controls

2.  Identifying which newly found variants are associated with SLE

3.  Selecting variants with a high likelihood of producing functional variants

4.  Verifying biological relevance of variants by *in vitro* study

5.  Verifying functional relevance in human subjects

## Developing analysis tools for massive amounts of sequencing data

The ability to cheaply and rapidly sequence large numbers of samples
has massively increased the amount of data processing required by
researchers. Compounding this increase in data, most existing tools
have not been designed to take full advantage of the current advances
in computer architecture, which parallel solutions to problems which
can be run on architectures with vastly differing computational
abilities (such as GPUs).

To resolve this, I am in the process of actively developing new open
source tools[^extending_note] which are capable of running on
massively parallel architectures using both multiple computers
(openMPI) and multiple GPUs on the same computer (Nvidia's CUDA) to

1.  Develop an open source massively parallel imputation method which
    can combine existing GWAS results with new deep sequencing
    and genetic profiling results.

2.  Extend same tools to call SNPs both incrementally and in parallel.

I am currently working on extending samtools (an Open Source SNP
calling suite) to support running on multiple computers.

## Gut microbiota alteration in SLE

Many autoimmune disorders are known to be affected by gut microbiota.
Preliminary evidence suggests that mice which develop SLE have
differences in gut microbiota from mice which do not develop SLE,
which suggests that SLE severity may also be affected by differences
in human gut microbiota. I am currently developing novel methods to

1.  Determine which gut microbiota differ between mice with and
    without SLE and using 16S sequencing

2.  Determine which gut microbiota differ between humans with and
    without SLE using 16S sequencing, given #1.

## Continuous, incremental analysis

Just as science requires continuous testing of hypotheses,
bioinformatics is slowly moving towards continuous incremental
analysis of data. Most current tools use iterative analysis, where
analyses must be completely re-run with each new piece of data that is
obtained. When the amount of data remains small relative to the
overall computing power available, this is a feasible approach.
However, as the amount of data increases, it stops being feasible to
completely reanalyze data as new data is obtained, and incremental
analysis approaches are necessary. I will be working to extend
existing analysis pipelines to handle the incremental analysis of
data.

# Research Funding

## Funding Opportunities

I anticipate obtaining funding from the following sources to pursue
the research goals outlined previously: 

1.  National Institute of Health
    
    1.  [Research Project Grant (RO1)](http://grants.nih.gov/grants/guide/pa-files/PA-13-302.html) (NHGRI, NIAID, BISTI)
    
    2.  [NLM Career Development Award in Biomedical Informatics (K01)](http://grants.nih.gov/grants/guide/pa-files/PAR-13-284.html)
    
    3.  [Continued Development and Maintenance of Software (R01)](http://grants.nih.gov/grants/guide/pa-files/PAR-11-028.html) (extending existing biofinformatics software)

2.  Alliance for Lupus Research
    
    1.  [Target Identification in Lupus](http://www.lupusresearch.org/news-and-events/press-releases/til.html)

3.  Institution specific funding

## Track Record of Funding

The projects that I am proposing have a strong track record of being
funded by both the NIH and the Alliance for Lupus Research.

# Wider impact of research agenda


My research will identify genetic variants and pathways underlying SLE
and other important human diseases, leading to better methodologies
for both the diagnosis and treatment of those diseases, and resulting
in significant decreases is patient morbidity and mortality.
Secondarily, the tools that I develop to effect this work will enable
researchers in other fields to more rapidly and cheaply identify
relevant factors for economically and environmentally important
phenotypes, such as pathogen-resistance in crops or
disease-susceptibility in thylocenes. Furthermore, as all of my tools
will be released under Open Source licenses, external researchers will
be able to build upon and improve my tools without being forced to
reinvent them.


[#Armstrong2009:sle_gi]: D L Armstrong et al. “Identification of new SLE-associated genes with a two-step
    Bayesian study design”. In: Genes Immun. 10.5 (July 2009), pp. 446–456. doi:
    [10.1038/gene.2009.38](http://dx.doi.org/10.1038/gene.2009.38).
    
[#Armstrong2008:function2gene]: Don L Armstrong, Chaim O Jacob, and Raphael Zidovetzki. “Function2Gene: a
    gene selection tool to increase the power of genetic association studies by utilizing
    public databases and expert knowledge”. In: BMC Bioinformatics 9 (2008), p. 311.
    doi: [10.1186/1471-2105-9-311](http://dx.doi.org/10.1186/1471-2105-9-311).
    
[#Jacob2009:sle_irak1]: Chaim O Jacob et al. “Identification of IRAK1 as a risk gene with critical role
    in the pathogenesis of systemic lupus erythematosus”. In: Proc. Natl. Acad. Sci.
    U.S.A. 106.15 (Apr. 2009), pp. 6256–6261. doi: [10.1073/pnas.0901181106](http://dx.doi.org/10.1073/pnas.0901181106).

[#Jacob2007:ar_lupus]: Chaim O Jacob et al. “Identification of novel susceptibility genes in childhood-
    onset systemic lupus erythematosus using a uniquely designed candidate gene
    pathway platform”. In: Arthritis Rheum. 56.12 (Dec. 2007), pp. 4164–4173. doi:
    [10.1002/art.23060](http://dx.doi.org/10.1002/art.23060).

[#Jacob2012:sle_ncf2]: Chaim O Jacob et al. “Lupus-associated causal mutation in neutrophil cytosolic
    factor 2 (NCF2) brings unique insights to the structure and function of NADPH
    oxidase”. In: Proc. Natl. Acad. Sci. U.S.A. 109.2 (Jan. 2012), pp. 59–67. doi:
       [10.1073/pnas.1113251108](http://dx.doi.org/10.1073/pnas.1113251108).

[#Stapleton2011:thbs1]: Christopher J Stapleton et al.
     “Thrombospondin-1 modulates the angiogenic phe- notype of human
     cerebral arteriovenous malformation endothelial cells”. In:
     Neurosurgery 68.5 (May 2011), pp. 1342–1353.
     doi: [10.1227/NEU.0b013e31820c0a68](http://dx.doi.org/10.1227/NEU.0b013e31820c0a68).

[survival_rate]: Only 4% of patients survive to 5 years after diagnosis

[avm_definition]: Direct artery to vein connection without an
    intervening capilary bed; leads to high pressure arterial flow in
    venous tissue and can lead to hemmorrhage and death.

[extending_note]: and extending existing open source tools where they
    exist