The
story of Darwin's finches
reports,
that Charles Darwin
had
no idea that the birds he collected during his
voyage to the Galapagos Archipelago, could be a
key to his future scientific achievements. Only
after he gave the birds of which Darwin had
thought were blackbirds, to the
ornithologist John Gould.
This expert found, that "gross-bills" and
finches were
in fact "a series of ground Finches which are so
peculiar" as to form "an entirely new group,
containing 12 species." Since Darwin at this time
had not labeled the finches by their island of
origin, only after gathering more specimens from
his travel travelers he was able to
establish that
the species were unique to islands. This supplied
him with crucial data to formulate his
theory
of evolution.
Apart from the scientific travels on the Beagle and the collaboration with Gould, two more general developments were underlying developments that helped Darwin to conceive his groundbreaking theory: The growing interest of the Victorian era for natural history and the increasingly systematic scientific collection of natural specimens and the subsequent availability of experts (such as John Gould) who had begun to gather great knowledge and expertise on the differences and commonalities of animal species from local natural history collections.
The "scientific framework" in which all this took place was much older and can be traced back to the first names people gave to animals and plants over the work of Carl Linneaus gave rise to the modern day taxonomy.
Stem cell research faces currently the challenge to be have to not only define but also predict properties of cell preparations in vitro and in vivo.
In contrast to in vitro cell cultures, animal and plant species do offer inherently a great way of "communicating" differences and commonalities between individual species and their "developmental potential": Their over all physical appearance and features, such as bones and feathers are a visual and three dimensional "interface" that can be easily understood by effective and powerful "pattern recognition" algorithms implemented in the human brain (Fig. 2A).
This is not the case for cell populations in vitro or in vivo. Especially cells in culture represent an man made biological artifacts that is inherently difficult to distinguish from each others. Even today, it has been estimated, that in reality HeLa cells are contaminating or actually have inwardly supplanted about 10 - 20% of all cell lines grown in labs worldwide.
The categorization of cell types based on their origin can be accomplished by means of identifying unique DNA sequences ("genetic fingerprint") this method though has its limitations when the cell type in question is pluri- or totipotent and the particular differentiation state has to be determined.
Apart from the scientific travels on the Beagle and the collaboration with Gould, two more general developments were underlying developments that helped Darwin to conceive his groundbreaking theory: The growing interest of the Victorian era for natural history and the increasingly systematic scientific collection of natural specimens and the subsequent availability of experts (such as John Gould) who had begun to gather great knowledge and expertise on the differences and commonalities of animal species from local natural history collections.
The "scientific framework" in which all this took place was much older and can be traced back to the first names people gave to animals and plants over the work of Carl Linneaus gave rise to the modern day taxonomy.
Stem cell research faces currently the challenge to be have to not only define but also predict properties of cell preparations in vitro and in vivo.
In contrast to in vitro cell cultures, animal and plant species do offer inherently a great way of "communicating" differences and commonalities between individual species and their "developmental potential": Their over all physical appearance and features, such as bones and feathers are a visual and three dimensional "interface" that can be easily understood by effective and powerful "pattern recognition" algorithms implemented in the human brain (Fig. 2A).
This is not the case for cell populations in vitro or in vivo. Especially cells in culture represent an man made biological artifacts that is inherently difficult to distinguish from each others. Even today, it has been estimated, that in reality HeLa cells are contaminating or actually have inwardly supplanted about 10 - 20% of all cell lines grown in labs worldwide.
The categorization of cell types based on their origin can be accomplished by means of identifying unique DNA sequences ("genetic fingerprint") this method though has its limitations when the cell type in question is pluri- or totipotent and the particular differentiation state has to be determined.
"Figure
2 | Comparative analysis of CaM expression in
finches. a, b, Geospiza group species displaying
distinct beak morphologies form a monophyletic
group. c, The differences in beak morphology are
skeletal. d, CaM is expressed in a strong
distal–ventral domain in the mesenchyme of the
upper beak prominence of the large cactus finch, G.
conirostris, somewhat lower levels in cactus finch,
G. scandens, and at significantly lower levels in
the large ground finch and medium ground finch, G.
magnirostris and G. fortis, respectively. Very low
levels of CaM were detected in the mesenchyme of G.
difficilis, G. fuliginosa and the basal warbler
finch Certhidea olivacea. CaM expression domains
are indicated with short arrows in d. Scale bar,
1mmin b. The molecular tree is from ref. 23; images
of skulls are from ref. 6, with permission from the
author."
Abzhanov, A. et al. Nature 442, 563–567 (2006)
Abzhanov, A. et al. Nature 442, 563–567 (2006)
"The
model is diagrammed as a network depicting how the
visible variables v1,...,vn
in the bottom layer of nodes are generated from the
hidden variables h1,...,
hr in the top layer of nodes. According to the
model, the visible variables vi are generated from
a probability distribution with mean
aWia
ha. In the network diagram, the influence of h
a on vi is represented by a connection with
strength Wia. In the application to facial
images, the visible variables are the image
pixels, whereas the hidden variables contain
the parts-based encoding. For fixed a, the
connection strengths W1a,...,Wna
constitute a specific basis image (right middle)
which is combined with other basis images to
represent a whole facial image (right bottom)." Lee
& Seung 1999
Nature 401, 788-791
aWia
ha. In the network diagram, the influence of h
a on vi is represented by a connection with
strength Wia. In the application to facial
images, the visible variables are the image
pixels, whereas the hidden variables contain
the parts-based encoding. For fixed a, the
connection strengths W1a,...,Wna
constitute a specific basis image (right middle)
which is combined with other basis images to
represent a whole facial image (right bottom)." Lee
& Seung 1999
Nature 401, 788-791
Genome
wide transcriptional profiles have been used in the
past 10 years for the identification of subgroups
and molecular mechanisms defining cancers.
Microarray data repositories (e.g. NCBI GEO, Fig 1B) might serve as storage places of "transcriptional phenotypes", maybe the virtual equivalent comparable to natural history collections (e.g. like the Darwin Center at the Natural History Museum, London, Fig 1A)
The problem remains, to extract biologically meaningful sample groups from multidimensional data for which no human sense exists (Fig.2B). Machine learning algorithms can be used to do this efficiently.
Microarray data repositories (e.g. NCBI GEO, Fig 1B) might serve as storage places of "transcriptional phenotypes", maybe the virtual equivalent comparable to natural history collections (e.g. like the Darwin Center at the Natural History Museum, London, Fig 1A)
The problem remains, to extract biologically meaningful sample groups from multidimensional data for which no human sense exists (Fig.2B). Machine learning algorithms can be used to do this efficiently.
Fig. 1A
Fig. 1B
The
information about molecular similarities can be
visualized efficiently by TreeMaps by taking
advantage of a summary statistics (sample measures
from the consensus clustering
framework)
that captures the proximity of a sample to a
specific cluster. This value between 1 (belongs
to a cluster) and 0 (does not share similarities
with one cluster) can be used to modify the size
of a sample tile (Fig. 4B).
This technique can visualize the global view of a cluster and other clusters/samples that have similar features encoded in their transcriptional profiles as well as display interactively information about each sample, thus enabling an intuitive access to very large clustering results with more than five biologically relevant sample clusters.
This technique can visualize the global view of a cluster and other clusters/samples that have similar features encoded in their transcriptional profiles as well as display interactively information about each sample, thus enabling an intuitive access to very large clustering results with more than five biologically relevant sample clusters.
Fig. 4A
Fig. 3B
Reference
and Context - collect and define
resolving ambiguity with a "majority vote"
Examples:
The
Consensus Clustering Algorithm-
Natural
history museu of Ole Worm, Museum Wormiani Historia
(Engraving)
From: Museum Wormianum Published: Elzevier, Leiden 1655 Source: Wellcome Library, London.
From: Museum Wormianum Published: Elzevier, Leiden 1655 Source: Wellcome Library, London.
Fig. 2A
Fig. 2B
