CNV_FunctionalAnnotation (All currently done in hg19)

Details to annotate CNV and apply prediction model

Required python3 dependencies: defaultdict, pybedtools, numpy, pandas, scipy

Citation

Han L, Zhao X, Benton ML, Perumal T, Collins RL, Hoffman, GE, Johnson JS, Sloofman L, Stone MJ, CommonMind Consortium, Brennand KJ, Brand H, Sieberts SK, Marenco S, Peters MA, Lipska BK, Roussos P, Capra JA, Talkowski M, Ruderfer DM. Functional annotation of rare structural variation in the human brain. Nature Communications 2020.

Downloading and generating necessary regulatory annotation files (should take < ~5 minutes)

To download/generate annotation files run `bin/make_annotation_files`.

Prediction models

The deletion and duplication models are saved in CNV-models.Rdata

Annotating CNV and applying models to generate regulatory disruption scores (will generate intermediate and results files in same directory as CNV file)

sh bin/annotate-cnv.sh [path where git repository was cloned] [path_to_CNV_file]

There is a test CNV file in test/test.cnv (see below). Only DEL or DUP are accepted types

id	chr	start	end	SVType
CNV1	2	133450000	133550000	DEL
CNV2	12	12150000	12880000	DEL
CNV3	7	6450000	6550000	DUP
CNV4	4	70450000	71260000	DUP
CNV5	18	5450000	7550000	DEL

Intermediate CNV by gene annotation file (*.gene.annot)

Output file has same columns as input file plus below:

SVLen: CNV length
gene: Ensembl gene name
trans: proportion of entire gene affected ExonProp: proportion of exonic sequence affected
pro.sum: proportion of promoter sequence affected
enh.sum: sum proportion of all enhancers sequence affected
inTad: within TAD (1) or not (0)
pred_exp: predicted z-score of expression sum: cumulative sum of ExonProp, pro.sum and enh.sum cre: sum of cis regulatory elements pro.sum and enh.sum gene.cnt: number of genes with ExonProp> 0

id	chr	start	end	SVType	SVLen	gene	trans	ExonProp	pro.sum	enh.sum	inTad	pred_exp	sum	cre	gene.cnt
CNV1	2	133450000	133550000	DEL	100000	ENSG00000176771	0.111525	0.50330015715034	0	0	1	-0.923467521409079	0.50330015715034	0	1
CNV2	12	12150000	12880000	DEL	730000	ENSG00000070018	1	1	1	0	1	-2.11313705930251	2	1	1
CNV2	12	12150000	12880000	DEL	730000	ENSG00000111261	1	1	1	3	0	-2.14977734794863	5	4	1
CNV2	12	12150000	12880000	DEL	730000	ENSG00000111266	1	1	1	0	1	-2.11313705930251	2	1	1
CNV2	12	12150000	12880000	DEL	730000	ENSG00000111269	1	1	1	0	1	-2.11313705930251	2	1	1
CNV2	12	12150000	12880000	DEL	730000	ENSG00000111276	1	1	1	6	1	-2.20448655339361	8	7	1
CNV2	12	12150000	12880000	DEL	730000	ENSG00000121380	1	1	1	5	1	-2.18926163771176	7	6	1
CNV2	12	12150000	12880000	DEL	730000	ENSG00000165714	1	1	1	0	1	-2.11313705930251	2	1	1
CNV2	12	12150000	12880000	DEL	730000	ENSG00000178878	0.0110419	0.0354825348612453	1	0	1	-0.39996407327444	1.03548253486125	1	1

Final output file (*.reg_disruption)

Additional columns from input file:

ngenes: Number of genes affected directly by CNV
exon: Cumulative sum of proportion of exonic sequence affected across all genes
enh: Cumuatlve sum of all enhancer sequences affected across all genes
pro: Cumuatlve sum of all promoter sequences affected across all genes
pred_exp: Predicted expression z-score
reg_dist: Regulatory disruption score

id	chr	start	end	SVType	SVLen	ngenes	exon	enh	pro	pred_exp	reg_dist
CNV1	2	133450000	133550000	DEL	100000	1	0.50330015715034	0	0	-0.923467521409079	-0.683216839007316
CNV2	12	12150000	12880000	DEL	730000	22	21.0354825348612	36	22	-45.6165484017837	-14.6298854256054
CNV3	7	6450000	6550000	DUP	100000	5	4.34803581412534	21	4	4.90544957250558	1.67441995009775
CNV4	4	70450000	71260000	DUP	810000	21	20.4703337453646	0	21	30.8031235891722	9.62781209180071
CNV5	18	5450000	7550000	DEL	2100000	37	36.2654214123007	2	37	-88.1485171766695	-8.63831141801943