2022-03-22
051_ Analysis of schizophrenia patients using subgroup specific transcriptomic signatures
Research Question and Aims
The aim of this analysis is to explore whether signatures of schizophrenia subgroups, identified via subgroup analysis of gene expression data, are predictive of clinical course, treatment response, and clinical comorbidity. Using publicly available brain and blood gene expression data, we are applying subgroup identification strategies to identify reproducible signatures of patients with schizophrenia and are testing the association of these signatures with polygenic risk, as well as brain functional differences relevant for the illness. The employed subgroup identification methods integrate biological pathway information to increase their applicability on the high-dimensional data, and to provide improved insight into mechanisms relevant for divergent patient subgroups. We intend to use the deeply phenotyped PsyCourse data to validate the identified signatures, test association with course and treatment response, as well as their potential relevance for clinical comorbidity effects, in particular in relation to diabetes and cardiovascular diseases.
Analytic Plan
1. A biologically-informed subgroup identification algorithm was trained on publicly available prefrontal cortex gene expression data from patients with schizophrenia and controls. We have been using supervised machine learning in order to identify the underlying transcriptomic signature, in order to predict this signature into independent brain and blood validation data. Using paired gene expression and genetic association data, we are exploring associations between the predicted signature and polygenic risk for schizophrenia.
2. We intend to apply these pre-trained models to the Psycourse gene expression data and test whether a) the predicted scores related to divergent subgroups of patients with schizophrenia and bipolar disorder also in the PsyCourse data, b) there are associations between predicted scores and the clinical/phenotypic annotations, c) there are reproducible associations with polygenic risk for schizophrenia.
Resources needed
OMICS dataset: RNA-seq transcriptomic on case samples (Raw dataset already available through Riya and preprocessing pipeline will be from ZI/Heidelberg)
Clinical/phenotypic
Raw and imputed genotypes: GSA chip variables:
v1_clin:
v1_clin_med_medi_1
v1_clin_med_kategorie_1
v1_con:
v1_con_med_medi_2
v1_con_med_kategorie_2
Demographic:
v1_sex
v1_age
v1_marital_stat
v1_partner
v1_no_bio_chld
v1_liv_aln
v1_school
v1_prof_dgr
v1_ed_status
v1_curr_paid_empl
v1_disabl_pens
v1_spec_emp
v1_wrk_abs_pst_5_yrs
v1_cur_work_restr
Psychiatric history:
v1_cur_psy_trm
v1_outpat_psy_trm
v1_age_1st_out_trm
v1_daypat_inpat_trm
v1_age_1st_inpat_trm
v1_dur_illness
v1_1st_ep
v1_tms_daypat_outpat_trm
v1_cat_daypat_outpat_trm
Medication:
v1_adv
v1_medchange
v1_lith
v1_lith_prd (raw + summary variables)
Physical measures and somatic diseases:
v1_height
v1_weight
v1_waist
v1_bmi
v1_chol_trig
v1_hyperten
v1_ang_pec
v1_heart_att
v1_stroke
v1_diabetes
v1_hyperthy
v1_hypothy
v1_allerg
v1_neuroder
v1_psoriasis
v1_autoimm
v1_cancer
v1_kid_fail
v1_stone
v1_epilepsy
v1_migraine
v1_parkinson
v1_liv_cir_inf
v1_tbi
v1_beh
v1_eyear
v1_inf
Substance abuse:
v1_ever_smkd
v1_age_smk
v1_no_cig
v1_alc_pst12_mths
v1_lftm_alc_dep
v1_evr_ill_drg
DSM-IV Diagnosis:
v1_scid_dsm_dx
v1_scid_dsm_dx_cat
v1_scid_age_MDE
v1_scid_ever_delus
v1_scid_ever_halls
v1_scid_ever_psyc
Neuropsychology:
v1_nrpsy_lng
v1_nrpsy_mtv
v1_nrpsy_tmt_A_rt
v1_nrpsy_tmt_A_err
v1_nrpsy_tmt_B_rt
v1_nrpsy_tmt_B_err
v1_nrpsy_dgt_sp_frw
v1_nrpsy_dgt_sp_bck
v1_nrpsy_dg_sym
v1_nrpsy_mwtb
Symptom Rating Scales:
v1_panss_p1
v1_panss_p2
v1_panss_p3
v1_panss_p4
v1_panss_p5
v1_panss_p6
v1_panss_p7
v1_panss_sum_pos
v1_panss_n1
v1_panss_n2
v1_panss_n3
v1_panss_n4
v1_panss_n5
v1_panss_n6
v1_panss_n7
v1_panss_sum_neg
v1_panss_g1
v1_panss_g2
v1_panss_g3
v1_panss_g4
v1_panss_g5
v1_panss_g6
v1_panss_g7
v1_panss_g8
v1_panss_g9
v1_panss_g10
v1_panss_g11
v1_panss_g12
v1_panss_g13
v1_panss_g14
v1_panss_g15
v1_panss_g16
v1_panss_sum_gen
v1_panss_sum_tot
v1_idsc_itm1
v1_idsc_itm2
v1_idsc_itm3
v1_idsc_itm4
v1_idsc_itm5
v1_idsc_itm6
v1_idsc_itm7
v1_idsc_itm8
v1_idsc_itm9
v1_idsc_itm10
v1_idsc_itm11
v1_idsc_itm12
v1_idsc_itm13
v1_idsc_itm14
v1_idsc_itm15
v1_idsc_itm16
v1_idsc_itm17
v1_idsc_itm18
v1_idsc_itm19
v1_idsc_itm20
v1_idsc_itm21
v1_idsc_itm22
v1_idsc_itm23
v1_idsc_itm24
v1_idsc_itm25
v1_idsc_itm26
v1_idsc_itm27
v1_idsc_itm28
v1_idsc_itm29
v1_idsc_itm30
v1_idsc_sum
v1_ymrs_itm1
v1_ymrs_itm2
v1_ymrs_itm3
v1_ymrs_itm4
v1_ymrs_itm5
v1_ymrs_itm6
v1_ymrs_itm7
v1_ymrs_itm8
v1_ymrs_itm9
v1_ymrs_itm10
v1_ymrs_itm11
v1_ymrs_sum
v1_bdi2_itm1
v1_bdi2_itm2
v1_bdi2_itm3
v1_bdi2_itm4
v1_bdi2_itm5
v1_bdi2_itm6
v1_bdi2_itm7
v1_bdi2_itm8
v1_bdi2_itm9
v1_bdi2_itm10
v1_bdi2_itm11
v1_bdi2_itm12
v1_bdi2_itm13
v1_bdi2_itm14
v1_bdi2_itm15
v1_bdi2_itm16
v1_bdi2_itm17
v1_bdi2_itm18
v1_bdi2_itm19
v1_bdi2_itm20
v1_bdi2_itm21
v1_bdi2_sum
v1_cgi_s
v1_gaf