Skip to content

Commit

Permalink
Add more data types to incremental data upload folder
Browse files Browse the repository at this point in the history
  • Loading branch information
forus committed May 15, 2024
1 parent 911ae28 commit c7343f9
Show file tree
Hide file tree
Showing 11 changed files with 112 additions and 0 deletions.
15 changes: 15 additions & 0 deletions tests/system_tests_import_data.py
Original file line number Diff line number Diff line change
Expand Up @@ -107,6 +107,16 @@ def test_incremental_load(self, run_java, locate_jar):
'--meta', f'{data_directory}/meta_clinical_samples.txt', '--loadMode', 'bulkload', '--data', f'{data_directory}/data_clinical_samples.txt', '--noprogress')
mutation_call = call(*common_part, 'org.mskcc.cbio.portal.scripts.ImportProfileData', '--overwrite-existing',
'--meta', f'{data_directory}/meta_mutations_extended.txt', '--loadMode', 'bulkload', '--update-info', 'False', '--data', f'{data_directory}/data_mutations_extended.maf', '--noprogress')
cna_discrete_call = call(*common_part, 'org.mskcc.cbio.portal.scripts.ImportProfileData', '--overwrite-existing',
'--meta', f'{data_directory}/meta_cna_discrete.txt', '--loadMode', 'bulkload', '--update-info', 'False', '--data', f'{data_directory}/data_cna_discrete.txt', '--noprogress')
cna_log2_call = call(*common_part, 'org.mskcc.cbio.portal.scripts.ImportProfileData', '--overwrite-existing',
'--meta', f'{data_directory}/meta_cna_log2.txt', '--loadMode', 'bulkload', '--update-info', 'False', '--data', f'{data_directory}/data_cna_log2.txt', '--noprogress')
expression_median_call = call(*common_part, 'org.mskcc.cbio.portal.scripts.ImportProfileData', '--overwrite-existing',
'--meta', f'{data_directory}/meta_expression_median.txt', '--loadMode', 'bulkload', '--update-info', 'False', '--data', f'{data_directory}/data_expression_median.txt', '--noprogress')
methylation_hm27_call = call(*common_part, 'org.mskcc.cbio.portal.scripts.ImportProfileData', '--overwrite-existing',
'--meta', f'{data_directory}/meta_methylation_hm27.txt', '--loadMode', 'bulkload', '--update-info', 'False', '--data', f'{data_directory}/data_methylation_hm27.txt', '--noprogress')
treatment_ic50_call = call(*common_part, 'org.mskcc.cbio.portal.scripts.ImportProfileData', '--overwrite-existing',
'--meta', f'{data_directory}/meta_treatment_ic50.txt', '--loadMode', 'bulkload', '--update-info', 'False', '--data', f'{data_directory}/data_treatment_ic50.txt', '--noprogress')
case_list_call = call(*common_part, 'org.mskcc.cbio.portal.scripts.UpdateCaseListsSampleIds',
'--meta', f'{data_directory}/meta_clinical_samples.txt', '--case-lists', f'{data_directory}/case_lists')

Expand All @@ -115,6 +125,11 @@ def test_incremental_load(self, run_java, locate_jar):
clinical_patient_call,
clinical_sample_call,
mutation_call,
cna_discrete_call,
cna_log2_call,
expression_median_call,
methylation_hm27_call,
treatment_ic50_call,
case_list_call,
])

Expand Down
10 changes: 10 additions & 0 deletions tests/test_data/study_es_0_inc/data_cna_discrete.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
Hugo_Symbol Entrez_Gene_Id TCGA-A1-A0SB-01 TCGA-A1-A0SB-03 TCGA-BH-NEW
ACAP3 116983 0 0 -1
375790 -1 -1 0
ATAD3A 55210 0 0 -2
ATAD3B 83858 -2 -1 0
ATAD3C 219293 0 0 0
#AURKAIP1 54998
ERCC5 2073 0 -1 -2
ACP3 55 0 0 0
TP53 -1 0 -2
10 changes: 10 additions & 0 deletions tests/test_data/study_es_0_inc/data_cna_log2.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
Hugo_Symbol Entrez_Gene_Id TCGA-A1-A0SB-01 TCGA-A1-A0SB-03 TCGA-BH-NEW
ACAP3 116983 0.751 0.533 0.114
375790 0.062 0.071 0.948
ATAD3A 55210 0.487 0.695 0.364
ATAD3B 83858 0.150 0.492 0.300
ATAD3C 219293 0.995 0.170 0.654
#AURKAIP1 54998
ERCC5 2073 0.816 0.514 0.165
ACP3 55 0.252 0.713 0.513
TP53 0.360 0.538 0.891
10 changes: 10 additions & 0 deletions tests/test_data/study_es_0_inc/data_expression_median.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
Hugo_Symbol Entrez_Gene_Id TCGA-A1-A0SB-01 TCGA-A1-A0SB-03 TCGA-BH-NEW
ACAP3 116983 0.096 0.826 0.032
375790 0.309 0.399 0.680
ATAD3A 55210 0.569 0.189 0.266
ATAD3B 83858 0.829 0.473 0.611
ATAD3C 219293 0.307 0.445 0.045
#AURKAIP1 54998
ERCC5 2073 0.171 0.766 0.590
ACP3 55 0.422 0.870 0.745
TP53 0.179 0.694 0.808
10 changes: 10 additions & 0 deletions tests/test_data/study_es_0_inc/data_methylation_hm27.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
Hugo_Symbol Entrez_Gene_Id TCGA-A1-A0SB-01 TCGA-A1-A0SB-03 TCGA-BH-NEW
ACAP3 116983 0.022 0.681 0.790
375790 0.435 0.340 0.321
ATAD3A 55210 0.229 0.946 0.439
ATAD3B 83858 0.885 0.707 0.664
ATAD3C 219293 0.660 0.315 0.694
#AURKAIP1 54998
ERCC5 2073 0.436 0.749 0.345
ACP3 55 0.622 0.396 0.029
TP53 0.563 0.686 0.607
11 changes: 11 additions & 0 deletions tests/test_data/study_es_0_inc/data_treatment_ic50.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,11 @@
ENTITY_STABLE_ID NAME DESCRIPTION URL TCGA-A1-A0SB-01 TCGA-A1-A0SB-03 TCGA-BH-NEW
17-AAG Name of 17-AAG Desc of 17-AAG Url of 17-AAG 0.315 0.329701692 0.053038094
AEW541 Name of AEW541 Desc of AEW541 Url of AEW541 >8 2.353 2.68212986
AZD0530 Name of AZD0530 Desc of AZD0530 Url of AZD0530 0.234 >8 4.597949505
AZD6244 Name of AZD6244 Desc of AZD6244 Url of AZD6244 >8 >8 >8
Erlotinib Name of Erlotinib Desc of Erlotinib Url of Erlotinib >8 >8 >8
Irinotecan Name of Irinotecan Desc of Irinotecan Url of Irinotecan NA 0.083 NA
L-685458 Name of L-685458 Desc of L-685458 Url of L-685458 >8 >8 3.267752409
#Lapatinib Name of Lapatinib Desc of Lapatinib Url of Lapatinib
LBW242 Name of LBW242 Desc of LBW242 Url of LBW242 NA >8 >8
Nilotinib Name of Nilotinib Desc of Nilotinib Url of Nilotinib >8 >8 NA
10 changes: 10 additions & 0 deletions tests/test_data/study_es_0_inc/meta_cna_discrete.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
cancer_study_identifier: study_es_0
genetic_alteration_type: COPY_NUMBER_ALTERATION
datatype: DISCRETE
stable_id: gistic
show_profile_in_analysis_tab: true
profile_description: Putative copy-number from GISTIC 2.0. Values: -2 = homozygous deletion; -1 = hemizygous deletion; 0 = neutral / no change; 1 = gain; 2 = high level amplification.
profile_name: Putative copy-number alterations from GISTIC
data_filename: data_cna_discrete.txt
pd_annotations_filename: data_cna_pd_annotations.txt
namespaces: CustomNamespace
8 changes: 8 additions & 0 deletions tests/test_data/study_es_0_inc/meta_cna_log2.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,8 @@
cancer_study_identifier: study_es_0
genetic_alteration_type: COPY_NUMBER_ALTERATION
datatype: LOG2-VALUE
stable_id: log2CNA
show_profile_in_analysis_tab: false
profile_description: Log2 copy-number values for each gene (from Affymetrix SNP6).
profile_name: Log2 copy-number values
data_filename: data_cna_log2.txt
8 changes: 8 additions & 0 deletions tests/test_data/study_es_0_inc/meta_expression_median.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,8 @@
cancer_study_identifier: study_es_0
genetic_alteration_type: MRNA_EXPRESSION
datatype: CONTINUOUS
stable_id: mrna
profile_description: Expression levels (Agilent microarray).
show_profile_in_analysis_tab: false
profile_name: mRNA expression (microarray)
data_filename: data_expression_median.txt
8 changes: 8 additions & 0 deletions tests/test_data/study_es_0_inc/meta_methylation_hm27.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,8 @@
cancer_study_identifier: study_es_0
genetic_alteration_type: METHYLATION
datatype: CONTINUOUS
stable_id: methylation_hm27
profile_description: Methylation beta-values (HM27 platform). For genes with multiple methylation probes, the probe least correlated with expression is selected.
show_profile_in_analysis_tab: false
profile_name: Methylation (HM27)
data_filename: data_methylation_hm27.txt
12 changes: 12 additions & 0 deletions tests/test_data/study_es_0_inc/meta_treatment_ic50.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,12 @@
cancer_study_identifier: study_es_0
genetic_alteration_type: GENERIC_ASSAY
generic_assay_type: TREATMENT_RESPONSE
datatype: LIMIT-VALUE
stable_id: treatment_ic50
profile_name: IC50 values of compounds on cellular phenotype readout
profile_description: IC50 (compound concentration resulting in half maximal inhibition) of compounds on cellular phenotype readout of cultured mutant cell lines.
data_filename: data_treatment_ic50.txt
show_profile_in_analysis_tab: true
pivot_threshold_value: 0.1
value_sort_order: ASC
generic_entity_meta_properties: NAME,DESCRIPTION,URL

0 comments on commit c7343f9

Please sign in to comment.