AIBS02

1
- A computational Aid Tool for
Finding Enrichment in Various Data Sets
Tali Goren, Ohad Manor, Tommy Kaplan and Nir
Friedman
School of Computer Science Engineering, The
Hebrew University, Israel
FHL1-Summary
7
Problems in Molecular Biology
Yeast Regulation Network
4
1
(b)
(a)
(b)
Enriched TFs
RAP1 SFP1 GAT3

• A partial regulation network of Yeast cell was
  created using the method described at (3).
• Zooming into the network, we see that ribosomal
  related transcription factors were predicted with
  no false positives.
• In the process of Iron utilization, two TF are
  known to utilize iron, and the third is known to
  utilize Proline.

(c)
(c)

1. Using the previously described methods, one can
   learn about enriched experimental conditions
   along with FHL1. Growth conditions and stress
   response show enrichment.
2. A GO annotation which shows enrichment for FHL1
   is Ribosome assembly.
3. RAP1, SFP1 and GAT3 areTFs which are best
   correlated to FHL1 in the sense of genes
   regulated, all known to be highly involved in
   ribosome biogenesis.

1
Studying Heat shock Effect

8
Studying FHL1-Using Related TFs

5

Is the effect of a heat shock on peripheral
proteins different than the affect on
mitochondrial proteins?
What is ?
2
What is known about FHL1?

• Computational tool to check statistical
  enrichment
• Implemented in Perl
• Interactive command line
• May be scripted
• Concatenate tests and matrix operations
• Data manipulation functions and queries

• Putative transcriptional regulator
• Predicted to be involved in stress response
• Required for rRNA processing
• Null mutant shows reduced growth rate
• Can we discover these properties using
  ?

Binary Measurements
Non-Parametric Tests
Parametric Tests

1. ChIP analysis of gene binding to the FHL1
   transcription factor, and ChIP analysis of all
   the other TFs.
2. Hyper-Geometric test results in a matrix of
   p-values, connecting the FHL1 transcription
   factor to all the other TFs of Yeast. Enriched
   TFs resemble in genes regulated by them to FHL1
   (e.g. RAP1, GAT3 are similar and FKH2 is not)

• Gene expression in various experiments and
  Localization analysis of genes using GFP marking.
• (b) Unpaired T-test outputs a matrix of the
  resulting p-vals. A red cell represents a
  significant p-value, meaning that genes allocated
  to a certain cellular compartment show
  significant difference in expression than genes
  which do not have the same localization.
• (c) Conversion to a binary matrix using a
  threshold.
• (d) A binary values set for the same experiments,
  of various conditions involving a heat shock
  (e.g. temp., length, etc.)
• (e) Hyper-Geometric results in a matrix of
  p-vals, connecting the mitochondrial genes and
  the cell periphery genes to various heat shock
  conditions. The results suggest mit. genes they
  are more effected by heat shock than cell
  periphery genes.

Kolmogorov-Smirnov
Unpaired T test
Compare two unpaired groups
Chi-Square test
Wilcoxon test
Paired T test
Compare two paired groups
Spearman Correlation
Pearson Correlation
Quantify association between two variables
Studying FHL1-Experimental Conditions
Which Biological pathway involves a given TF?
6
Summary and Future Directions
9
3

References

• Joseph T. et al. The transcription factor Ifh1 is
  a key regulator of yeast ribosomal protein genes.
  Nature. 2004. 432 1054-1058.
• Dietmar E. Martin et al. TOR Regulates Ribosomal
  Protein Gene Expression via PKA and the Forkhead
  Transcription Factor FHL1.
• Cell. 2004. 119(7) 969-979.
• Audry P. Gasch et al. Genomic Expression Programs
  in the Response of Yeast Cells to Environmental
  Changes. Mol. Biol. Cell. 2000. 11(12) 4241
  4257.

1. Data sets of the Yeast genes gene expression in
   various experiments and ChIP of gene binding to
   FHL1 TF.
2. Unpaired T-test results in a matrix of p-vals,
   where each cell represents the resulting p-value
   for the corresponding vectors, which is converted
   to a binary matrix by a threshold.
3. Same expression experiments, regarding various
   conditions such as temperature, starvation,
   stress etc.
4. Hyper-Geometric test results in a matrix of
   p-vals, connecting FHL1 to various experimental
   conditions. Enriched conditions may be extracted
   (e.g. heat shock, invasive growth).

1. Data sets of the Yeast genes ChIP analysis of
   TF binding genes promoters and GO annotations.
2. Performing a Hyper-Geometric test results in a
   matrix of p-values.
3. Conversion to a binary matrix using a
   threshold.The result shows which TFs are enriched
   along with certain GO annotation and one can
   infer that this TF may be involved in that
   Biological process (e.g. RAP1, SFP1, FHL1 in
   ribosomal processes)