GSEA Settings

Notes:

for GSEA preRanked the ranked gene lists were created by choosing the probe-set for a gene that has a t-Statistic with the largest deviation from 0, resulting in a table with the format:
GENE_SYMBOL{tab}T-STATISTIC
Positive values of the t-Statistic are related to a high expression for the first phenotype compared to the second and negative values exactly the opposite.
for GSEA-"tTest" and GSEA-"log2_Ratio_of_Classes" the expression matrices were collapsed externally with the program collapse_ExpressionMatrix.py using an algorithm identical to the max_probe setting in GSEA.
(see table "max_probe" in the GSEA Manual)
the expression matrices stored in EnrichmentMap Session files for the preRanked "moderated t-Test" analyses contain the expression values from the same probe-set that was chosen to represent a gene in the ranked gene list.
the expression matrices stored in EnrichmentMap Session files for the GSEA-"tTest" and GSEA-"log2_Ratio_of_Classes" analyses contain for that expression value for a certain gene, that was the highest from all probe-sets for this gene in a given sample.
(see table "max_probe" in the GSEA Manual)

GSEA-settings for preRanked with "moderated t-Test":

Name of Option in GUI mode	command line option	value	Info
Required Fields
Gene sets database	-gmx	${GENE_SET}.gmt	respective gmt file
Number of permutations	-nperm	1000
Ranked List	-rnk	${PHENOTYPE}.rnk	respective ranked gene list
Collapse dataset to gene symbols	-collapse	no	collapsing was done externally with collapse_ExpressionMatrix.py
Chip platforms	-chip	not used	due to "-collapse no"
Basic Fields
Analysis name	-rpt_label	${SET}_${PHENOTYPE}__${GENE_SET}
Enrichment statistic	-scoring_scheme	weighted
Max size: exclude larger sets	-set_max	500
Min size: exclude smaller sets	-set_min	10
Save results in folder:	-out	${WORK_DIR}/${SET}/gsea_results
Advanced Fields
Collapsing mode for probe sets => 1 gene	-mode	not used	due to "-collapse no"
Normalisation mode	-norm	meandiv
Omit features with no Symbol	-include_only_symbols	true
Make detailed geneset report	-make_sets	true
Plot graphs for the top sets of each Phenotype	-plot_top_x	20
Seed for permutation	-rnd_seed	timestamp
Make a zipped file with all reports	-zip_report	false

Name of Option in GUI mode

command line option

value

Info

Required Fields

Gene sets database

-gmx

${GENE_SET}.gmt

respective gmt file

Number of permutations

-nperm

1000

Ranked List

-rnk

${PHENOTYPE}.rnk

respective ranked gene list

Collapse dataset to gene symbols

-collapse

collapsing was done externally with collapse_ExpressionMatrix.py

Chip platforms

-chip

not used

due to "-collapse no"

Basic Fields

Analysis name

-rpt_label

${SET}_${PHENOTYPE}__${GENE_SET}

Enrichment statistic

-scoring_scheme

weighted

Max size: exclude larger sets

-set_max

500

Min size: exclude smaller sets

-set_min

Save results in folder:

-out

${WORK_DIR}/${SET}/gsea_results

Advanced Fields

Collapsing mode for probe sets => 1 gene

-mode

not used

due to "-collapse no"

Normalisation mode

-norm

meandiv

Omit features with no Symbol

-include_only_symbols

true

Make detailed geneset report

-make_sets

true

Plot graphs for the top sets of each Phenotype

-plot_top_x

Seed for permutation

-rnd_seed

timestamp

Make a zipped file with all reports

-zip_report

false

GSEA-setting for "GSEA t-Test":

Option (in GUI mode)	command line option	value	Info
Required Fields
Expression dataset	-res	${GSEA_GCT}.gct	respective expression table
Gene sets database	-gmx	${GENE_SET}.gmt	respective gmt file
Number of permutations	-nperm	1000
Phenotype Labels	-cls	${PHENOTYPE_FILE}.cls#${PHENOTYPE}	respective ranked gene list
Collapse dataset to gene symbols	-collapse	no	collapsing was done externally with collapse_ExpressionMatrix.py
Permutation type	-permute	gene_set
Chip platforms	-chip	not used	due to "-collapse no"
Basic Fields
Analysis name	-rpt_label	${SET}_${PHENOTYPE}__${GENE_SET}
Enrichment statistic	-scoring_scheme	weighted
Metric for ranking genes	-metric	tTest
Gene list sorting mode	-sort	real
Gene list ordering mode	-order	descending
Max size: exclude larger sets	-set_max	500
Min size: exclude smaller sets	-set_min	10
Save results in folder:	-out	${WORK_DIR}/${SET}/gsea_results
Advanced Fields
Collapsing mode for probe sets => 1 gene	-mode	not used	due to "-collapse no"
Normalisation mode	-norm	meandiv
Randomization mode	-rnd_type	no_balance
Omit features with no Symbol	-include_only_symbols	true
Make detailed geneset report	-make_sets	true
Median for class metrics	-median	false
Number of Markers	-num	100
Plot graphs for the top sets of each Phenotype	-plot_top_x	20
Seed for permutation	-rnd_seed	timestamp
Make a zipped file with all reports	-zip_report	false

Option (in GUI mode)

command line option

value

Info

Required Fields

Expression dataset

-res

${GSEA_GCT}.gct

respective expression table

Gene sets database

-gmx

${GENE_SET}.gmt

respective gmt file

Number of permutations

-nperm

1000

Phenotype Labels

-cls

${PHENOTYPE_FILE}.cls#${PHENOTYPE}

respective ranked gene list

Collapse dataset to gene symbols

-collapse

collapsing was done externally with collapse_ExpressionMatrix.py

Permutation type

-permute

gene_set

Chip platforms

-chip

not used

due to "-collapse no"

Basic Fields

Analysis name

-rpt_label

${SET}_${PHENOTYPE}__${GENE_SET}

Enrichment statistic

-scoring_scheme

weighted

Metric for ranking genes

-metric

tTest

Gene list sorting mode

-sort

real

Gene list ordering mode

-order

descending

Max size: exclude larger sets

-set_max

500

Min size: exclude smaller sets

-set_min

Save results in folder:

-out

${WORK_DIR}/${SET}/gsea_results

Advanced Fields

Collapsing mode for probe sets => 1 gene

-mode

not used

due to "-collapse no"

Normalisation mode

-norm

meandiv

Randomization mode

-rnd_type

no_balance

Omit features with no Symbol

-include_only_symbols

true

Make detailed geneset report

-make_sets

true

Median for class metrics

-median

false

Number of Markers

-num

100

Plot graphs for the top sets of each Phenotype

-plot_top_x

Seed for permutation

-rnd_seed

timestamp

Make a zipped file with all reports

-zip_report

false

GSEA reports for "Log2_Ratio_of_Classes":

Option (in GUI mode)	command line option	value	Info
Required Fields
Expression dataset	-res	${GSEA_GCT}.gct	respective expression table
Gene sets database	-gmx	${GENE_SET}.gmt	respective gmt file
Number of permutations	-nperm	1000
Phenotype Labels	-cls	${PHENOTYPE_FILE}.cls#${PHENOTYPE}	respective ranked gene list
Collapse dataset to gene symbols	-collapse	no	collapsing was done externally with collapse_ExpressionMatrix.py
Permutation type	-permute	gene_set
Chip platforms	-chip	not used	due to "-collapse no"
Basic Fields
Analysis name	-rpt_label	${SET}_${PHENOTYPE}__${GENE_SET}
Enrichment statistic	-scoring_scheme	weighted
Metric for ranking genes	-metric	log2_Ratio_of_Classes
Gene list sorting mode	-sort	real
Gene list ordering mode	-order	descending
Max size: exclude larger sets	-set_max	500
Min size: exclude smaller sets	-set_min	10
Save results in folder:	-out	${WORK_DIR}/${SET}/gsea_results
Advanced Fields
Collapsing mode for probe sets => 1 gene	-mode	not used	due to "-collapse no"
Normalisation mode	-norm	meandiv
Randomization mode	-rnd_type	no_balance
Omit features with no Symbol	-include_only_symbols	true
Make detailed geneset report	-make_sets	true
Median for class metrics	-median	false
Number of Markers	-num	100
Plot graphs for the top sets of each Phenotype	-plot_top_x	20
Seed for permutation	-rnd_seed	timestamp
Make a zipped file with all reports	-zip_report	false

Option (in GUI mode)

command line option

value

Info

Required Fields

Expression dataset

-res

${GSEA_GCT}.gct

respective expression table

Gene sets database

-gmx

${GENE_SET}.gmt

respective gmt file

Number of permutations

-nperm

1000

Phenotype Labels

-cls

${PHENOTYPE_FILE}.cls#${PHENOTYPE}

respective ranked gene list

Collapse dataset to gene symbols

-collapse

collapsing was done externally with collapse_ExpressionMatrix.py

Permutation type

-permute

gene_set

Chip platforms

-chip

not used

due to "-collapse no"

Basic Fields

Analysis name

-rpt_label

${SET}_${PHENOTYPE}__${GENE_SET}

Enrichment statistic

-scoring_scheme

weighted

Metric for ranking genes

-metric

log2_Ratio_of_Classes

Gene list sorting mode

-sort

real

Gene list ordering mode

-order

descending

Max size: exclude larger sets

-set_max

500

Min size: exclude smaller sets

-set_min

Save results in folder:

-out

${WORK_DIR}/${SET}/gsea_results

Advanced Fields

Collapsing mode for probe sets => 1 gene

-mode

not used

due to "-collapse no"

Normalisation mode

-norm

meandiv

Randomization mode

-rnd_type

no_balance

Omit features with no Symbol

-include_only_symbols

true

Make detailed geneset report

-make_sets

true

Median for class metrics

-median

false

Number of Markers

-num

100

Plot graphs for the top sets of each Phenotype

-plot_top_x

Seed for permutation

-rnd_seed

timestamp

Make a zipped file with all reports

-zip_report

false