Annual Scientific Report 2015

Teichmann Group
Gene Expression Regulation and
Protein Complex Assembly
Our group seeks to elucidate general principles of gene expression and protein
complex assembly. We study protein complexes in terms of their 3D structure,
structural evolution and the principles underlying protein-complex formation
and organisation.
Our group seeks to elucidate general principles of
gene expression and protein complex assembly. We
endeavour to understand how changes in cell state are
regulated at the transcriptomic and epigenetic levels
by studying the differentiation of mouse T-helper (Th)
cells and embryonic stem cells (mESC) at the single-cell
level. We develop novel computational and experimental
approaches in single cell-genomics to address our
questions. We also study the 3D structure and evolution
of protein complexes, and the principles underlying
protein-complex formation and organisation.
The recent development of computational and
experimental tools in single cell genomics has led
us to many exciting new discoveries in single-cell
biology. We focus on the evolution and dynamics of
regulatory and physical interaction networks, combining
computational and mathematical approaches with
genome-wide and gene/protein experiments.
Major achievements
In 2015 the field of single-cell genomics expanded
rapidly, allowing our group to exploit scRNA-seq
technology to make many exciting new discoveries in
both stem cells and T cells. It was an exciting time of
change as Sarah was awarded the European Molecular
Biology Organisation (EMBO) Gold Award and our
group prepared to move to the Wellcome Trust
Sanger Institute.
Transcriptional regulation is critical for maintaining
the pluripotency of mESCs, and culture conditions
are also important for their self-renewal in vitro.
Using single-cell RNA-seq (scRNA-seq) approach, we
investigated the transcriptome profiles of mESCs in
three different culture conditions (serum, 2i, and a2i),
which represent slightly different pluripotent states.
The past understanding of these distinct pluripotent
states is limited to “bulk” analyses, which fail to capture
the complexity of cellular states within mESCs. Our
work (Kolodziejczyk et al., 2015) revealed additional
pluripotency network genes, including Ptma and Zfp640,
which demonstrate the value of this resource for
future discovery.
We use single-cell technology to study immune-cell
development, notably the T-cell receptor (TCR) – a
diverse protein mediating the recognition of antigen
fragments as peptides bound to major histocompatibility
complex (MHC) molecules. To investigate TCR
clonal relationships between cells alongside their
transcriptional profiles and functional responses, we
developed a novel computational method, TraCeR
(Stubbington, Lonnberg et al., 2016). This new method
enables us to link TCR sequence with transcriptional
profiles in individual cells with high accuracy and
sensitivity. By applying TraCeR to scRNA-seq data from
a mouse Salmonella infection model, we showed that
T-cell clonotypes span early-activated CD4+ T cells as
well as mature effector and memory cells.
One of the key challenges in single-cell genomics is
to ensure that only single, live cells are included in
downstream analysis. With the aim of increasing data
quality and bioinformatics reliability, we developed a
computational pipeline (Ilicic et al., 2016) for processing
scRNA-seq data and filtering out low-quality cells using
a curated set of over 20 biological and technical features.
Our pipeline ensures that only correctly annotated cells
are included in analyses – a crucial tool for drawing
more accurate biological conclusions.
We used mass spectrometry data together with a
large-scale analysis of structures of protein complexes
to examine the fundamental steps of protein assembly.
In collaboration with colleagues at the Cavendish
Laboratory at the University of Cambridge, we
analysed the tens of thousands of protein complexes
for which three-dimensional structures have been
experimentally determined, and identified repeating
patterns in the assembly transitions that occur (Ahnert
et al., 2015). We used these patterns to create a new
‘Periodic Table’ of protein complexes, which provides a
predictive framework for anticipating new, unobserved
topologies of protein complexes. The core work for
this study is in theoretical physics and computational
biology, combined with mass spectrometry work by our
colleagues at Oxford University.
145
2015 EMBL-EBI Annual Scientific Report