population of various types of cells, therefore creating an average
transcriptomic profile of the tissue. This can become an issue when
rare cell types are of interest, because their signal is essentially lost in
the noise and more abundant transcripts. One way to get around
this issue is by enriching for the population of interest, using a cell
surface marker, fluorescence or antibody, however, this will still
provide an averaged transcriptome across cells and does not capture
heterogeneity at the single cell level. Another way to improve the
analysis is to perform a temporal trajectory of the material in
question. For embryonic material this can be highly informative
because it adds the element of time, although still, within each
sample the heterogeneity produces noise.
Single cell RNA-sequencing has the potential to eliminate
much of the noise within a mixed population of cells. With a
temporal profile it enables the investigator to probe the transcriptional dynamics of heterogeneous cell populations because it measures the distribution of mRNA expression from individual cells.
Single cell transcriptomes can be profiled for a number of purposes
such as creating cell atlases, mapping cell lineage trajectories [4–
10], modeling virtual in situ hybridization [11] and more
[12]. Using scRNA, one can capture cell trajectories and developmental processes such as an EMT by applying a scRNA-seq timecourse to construct a cell trajectory map [13]. Generating an EMT
time-course to capture transient cell states at single cell resolution
informs the investigator with information on how this dynamic
process occurs over time, providing a resource that is not available
in any other known way.
Single cell RNA-sequencing is rapidly becoming a viable alternative to bulk RNA-sequencing, however, there are still some
inherent issues with the platform. One challenge is due to the fact
that RNA is harvested from only a single cell, and generally needs to
be amplified with reverse transcription or PCR. This process of
amplification can introduce bias that can lead to an incorrect interpretation. However, this can be overcome during the normalization
and computational analysis by using Unique Molecular Identifiers
(UMI), to uniquely label individual RNA molecules, greatly reducing amplification bias. Additionally, due to the sparsity of some
RNA transcripts present in the cell and the inefficient cell capture
process, sometimes a gene may have moderate expression in some
cells, but cannot be detected in another cell. These occurrences,
known as gene dropouts can be misleading because it is difficult to
differentiate between inefficiency of transcript capture and a cell
lacking that particular gene expression, or a gene that is expressed
intermittently, therefore dimensionality reduction and normalization should to be performed computationally [14, 15].
306
Abdull J. Massri et al.
transcriptomic profile of the tissue. This can become an issue when
rare cell types are of interest, because their signal is essentially lost in
the noise and more abundant transcripts. One way to get around
this issue is by enriching for the population of interest, using a cell
surface marker, fluorescence or antibody, however, this will still
provide an averaged transcriptome across cells and does not capture
heterogeneity at the single cell level. Another way to improve the
analysis is to perform a temporal trajectory of the material in
question. For embryonic material this can be highly informative
because it adds the element of time, although still, within each
sample the heterogeneity produces noise.
Single cell RNA-sequencing has the potential to eliminate
much of the noise within a mixed population of cells. With a
temporal profile it enables the investigator to probe the transcriptional dynamics of heterogeneous cell populations because it measures the distribution of mRNA expression from individual cells.
Single cell transcriptomes can be profiled for a number of purposes
such as creating cell atlases, mapping cell lineage trajectories [4–
10], modeling virtual in situ hybridization [11] and more
[12]. Using scRNA, one can capture cell trajectories and developmental processes such as an EMT by applying a scRNA-seq timecourse to construct a cell trajectory map [13]. Generating an EMT
time-course to capture transient cell states at single cell resolution
informs the investigator with information on how this dynamic
process occurs over time, providing a resource that is not available
in any other known way.
Single cell RNA-sequencing is rapidly becoming a viable alternative to bulk RNA-sequencing, however, there are still some
inherent issues with the platform. One challenge is due to the fact
that RNA is harvested from only a single cell, and generally needs to
be amplified with reverse transcription or PCR. This process of
amplification can introduce bias that can lead to an incorrect interpretation. However, this can be overcome during the normalization
and computational analysis by using Unique Molecular Identifiers
(UMI), to uniquely label individual RNA molecules, greatly reducing amplification bias. Additionally, due to the sparsity of some
RNA transcripts present in the cell and the inefficient cell capture
process, sometimes a gene may have moderate expression in some
cells, but cannot be detected in another cell. These occurrences,
known as gene dropouts can be misleading because it is difficult to
differentiate between inefficiency of transcript capture and a cell
lacking that particular gene expression, or a gene that is expressed
intermittently, therefore dimensionality reduction and normalization should to be performed computationally [14, 15].
306
Abdull J. Massri et al.