3. http://bar.utoronto.ca/~nprovart/unionDNaseHypersensitive
Sites.gff3
4. Older microarray platforms are able to detect varying numbers
of transcripts. Data from these are still quite useful and vast
numbers of expression profiling experiments have been conducted with them. The ATH1 array from Affymetrix has probe
sets for 22,814 transcripts, some of which may come from
several genes. Next-generation sequencing technologies, i.e.,
RNA-seq, are more comprehensive.
5. The Arabidopsis Genome Initiative identifier, AGI ID, is easily
found at TAIR.
6. It is useful to set the Signal Threshold to some value when
comparing different genes or viewing a number of different
Data Sources. That way, the expression level that “red” denotes
is constant. The expression-level distribution graph is also a
handy feature for determining if one’s gene of interest has a
strong level of expression. The small graph shows the distribution of the average expression level of all genes in the tissues
depicted on the output, while the red line shows where the
maximum expression level of the gene of interest falls along
that distribution.
7. The eFP-Seq Browser paper by O’Sullivan et al. has been
accepted at the Plant Journal.
8. The Bio-Analytic Resource does provide a bulk query tool
called
“Expression
Browser”
which
provides
a
Genevestigator-like ability to query many genes at once, see
http://bar.utoronto.ca/affydb/cgi-bin/affy_db_exprss_
browser_in.cgi.
9. Genevestigator has no control over experimental design, only a
post hoc analysis is possible to check the quality of the array.
For more information about quality control criteria visit
https://genevestigator.com/gv/file/GENEVESTIGATOR_
UserManual.pdf.
10. For experimental normalization, Genevestigator uses Bioconductor’s RMA implementation.
11. A p-value under 0.06 indicates that the signal is reliably
detected.
12. It is often useful to examine condition-dependent data sets, as
genes may respond one way in a set of tissues and in an opposite
way in others. If one lumps these sets together, then these
correlations cannot be detected. This issue is described in
greater detail in the Usadel et al. (2009) review [8].
13. Given the number of samples in most of these data sets, even a
Pearson correlation coefficient of 0.3 can be considered “significant.” But with this r-value, only (0.3)2 ¼ 9% of the
84
G. Alex Mason et al.
Sites.gff3
4. Older microarray platforms are able to detect varying numbers
of transcripts. Data from these are still quite useful and vast
numbers of expression profiling experiments have been conducted with them. The ATH1 array from Affymetrix has probe
sets for 22,814 transcripts, some of which may come from
several genes. Next-generation sequencing technologies, i.e.,
RNA-seq, are more comprehensive.
5. The Arabidopsis Genome Initiative identifier, AGI ID, is easily
found at TAIR.
6. It is useful to set the Signal Threshold to some value when
comparing different genes or viewing a number of different
Data Sources. That way, the expression level that “red” denotes
is constant. The expression-level distribution graph is also a
handy feature for determining if one’s gene of interest has a
strong level of expression. The small graph shows the distribution of the average expression level of all genes in the tissues
depicted on the output, while the red line shows where the
maximum expression level of the gene of interest falls along
that distribution.
7. The eFP-Seq Browser paper by O’Sullivan et al. has been
accepted at the Plant Journal.
8. The Bio-Analytic Resource does provide a bulk query tool
called
“Expression
Browser”
which
provides
a
Genevestigator-like ability to query many genes at once, see
http://bar.utoronto.ca/affydb/cgi-bin/affy_db_exprss_
browser_in.cgi.
9. Genevestigator has no control over experimental design, only a
post hoc analysis is possible to check the quality of the array.
For more information about quality control criteria visit
https://genevestigator.com/gv/file/GENEVESTIGATOR_
UserManual.pdf.
10. For experimental normalization, Genevestigator uses Bioconductor’s RMA implementation.
11. A p-value under 0.06 indicates that the signal is reliably
detected.
12. It is often useful to examine condition-dependent data sets, as
genes may respond one way in a set of tissues and in an opposite
way in others. If one lumps these sets together, then these
correlations cannot be detected. This issue is described in
greater detail in the Usadel et al. (2009) review [8].
13. Given the number of samples in most of these data sets, even a
Pearson correlation coefficient of 0.3 can be considered “significant.” But with this r-value, only (0.3)2 ¼ 9% of the
84
G. Alex Mason et al.