
Search BLAST Feedback Help

Library Statistics

GO Terms For Each Library

GO terms from each of the three ontologies (Biological Process, Molecular Function and Cellular Component) were assigned to many of the clusters in SCDb. We aggregated these categories into groupings that we consider biologically relevant in stem cell biology. Each bar represents the percent of the total number of clusters that have a given GO term associated.

Redundancy Analysis

Using Poisson statistics, we have analysed the redundancy of clusters in SCDb. Fitting our data to the Poisson distribution:

P(k)=(e^-mm^k)/k!

where k is the number of times a transcript is found and m is the Poisson mean. If we assume an equal representation of all transcripts within the libraries, we have sequenced about 23% of the expression space.

Clone Overlap in SCDb

Clicking on an area in the following Venn diagrams returns the list of unique clones it represents.

We compared non-singleton unique groups from mouse fetal liver and two separate bone marrow purification procedures. There is a large degree of overlap between all three libraries, as well as in the pairwise comparisons.

This diagram shows the comparison of mouse fetal liver, mouse bone marrow and human bone marrow. Again, there is a large degree of overlap between the the three sources, with a larger degree of overlap between the bone marrow samples than with the fetal liver samples. This is important because the mouse stem cell compartment is much better characterized than the human equivalent.

These diagrams show the comparison of non-singleton unique groups from RDA subtraction libraries where the tester and driver libraries were reversed. A comparison of the overlap between reversed subtractions should be a measure of the inherent noise in subtraction. The Rholo(-)Rhohi subtraction shows a much greater overlap on a gene-by-gene basis than the Sca+(-)Sca- subtraction. This is probably because the bone marrow populations being compared are much more similar than the fetal liver populations.

SCDb contains libraries using different PCR-based techniques for both mouse fetal liver and bone marrow. We can use this to compare libraries made using the different techniques. In this case, we show non-singleton unique groups. The genes represented in the RDA-based subtraction libraries show a large overlap with the corresponding capfinder-based libraries (43% for fetal liver SCA+(-)AA4.1-, 42% for bone marrow Rholo(-)Lin+).

We compared non-singleton unique groups from mouse fetal liver and two separate bone marrow purification procedures. There is a large degree of overlap between all three libraries, as well as in the pairwise comparisons.	This diagram shows the comparison of mouse fetal liver, mouse bone marrow and human bone marrow. Again, there is a large degree of overlap between the the three sources, with a larger degree of overlap between the bone marrow samples than with the fetal liver samples. This is important because the mouse stem cell compartment is much better characterized than the human equivalent.

These diagrams show the comparison of non-singleton unique groups from RDA subtraction libraries where the tester and driver libraries were reversed. A comparison of the overlap between reversed subtractions should be a measure of the inherent noise in subtraction. The Rholo(-)Rhohi subtraction shows a much greater overlap on a gene-by-gene basis than the Sca+(-)Sca- subtraction. This is probably because the bone marrow populations being compared are much more similar than the fetal liver populations.

SCDb contains libraries using different PCR-based techniques for both mouse fetal liver and bone marrow. We can use this to compare libraries made using the different techniques. In this case, we show non-singleton unique groups. The genes represented in the RDA-based subtraction libraries show a large overlap with the corresponding capfinder-based libraries (43% for fetal liver SCA+(-)AA4.1-, 42% for bone marrow Rholo(-)Lin+).