Pamgene phosphorylation chip analysis

Custom homebrew solution to analyse data from Pamgene kinase activity profiling system, created due to issues with the original Bionavigator software. Workflow was reverse engineered and approximated from various presentations and posters found online and as such is less complex and complete then the one offered commercially.

Example usage

pamgene.R -i raw_data_stk.csv --pamgene pamgene.csv --phosphonet phosphonet.csv --ctrl A1 --exp A2 --kinexus_score 300 --batch_correction

Commandline arguments

-i, --file_in input file containg data from pamgene kinase activity profiling system. See preview of input formats for more information.

-o, --dir_out output directory

--pamgene csv file containing information about peptides on the pamgene chip

--phosphonet csv file containing information about kinases phosphorylating specific protein residues of proteins included im pamgene csv file

-c, --ctrl control group name as a single string or comma-separated indices of alphabetically arranged <group_name>_<chip_id> denominators

-e, --exp experimental group name as a single string or comma-separated indices of alphabetically arranged <group_name>_<chip_id> denominators

--top_kinases top kinases from phosphonet file to include in calculations, if NULL the number will be estimated from data

--kinexus_score threshold kinexus version 2 score that determines whether the residue is considered as being phosphorylated by a kinase. All phosphorylations below threshold are omitted.

--no_perm number of permutations used to calculate specificity and selectivity scores

-b, --batch_correction whether to remove batch effect from peptide kinetics matrix. Uses sva::ComBat for calculations.

Example output

Peptide statistics

id	mean_ctrl	mean_exp	statistic	pvalue	A1_710250713	A1_710250715	A2_710250713	A2_710250715
ACM1_421_433	6.79	6.38	0.65	0.55	6.54	6.32	6.54	6.76
ACM1_444_456	6.70	6.46	0.37	0.74	6.53	6.42	6.85	7.02
ACM4_456_468	6.44	6.03	0.69	0.52	6.21	5.87	6.40	6.31
ACM5_494_506	11.02	10.70	0.22	0.84	11.39	10.67	11.06	11.62

Kinase scores

kinase_id	kinase_name	specificity	selectivity	total	sum_score	mean_score	sd_score
Q9UBS0	p70S6Kb (RPS6KB2)	0.70	1.00	1.70	-0.28	-0.01	0.03
P41743	PKCi (PRKCI)	0.36	1.00	1.36	-0.66	-0.33	0.48
P22694	PKACb PRKACB	1.00	0.34	1.34	-0.30	-0.01	0.02
P25098	BARK1 (ADRBK1; GRK2)	0.20	1.00	1.20	-0.45	-0.23	0.42

Preview of input formats

Raw data

Csv file with one of the following structures:

ID	UniprotAccession	<group1>_<chip_id1>_<time1>	<group1>_<chip_id1>_<time2>	etc.

for all groups, chips and time points, or same data transformed into long format with exactly 6 columns:

ID	UniprotAccession	group	chip_id	time	value

Pamgene

File describing the peptides included on the pamgene chip. Information can be inferred from the peptides reported in the raw data file by aligning the peptide protein sequences and adding proteins that pass similarity threshold and contain phosphoryltable amino acids in the corresponding frequences. Following columns are required in any order.

id	uniprot_id	phosphosite
ATF2_47_59	P15336	69
ATF2_47_59	P15336	71
ATF2_47_59	P15336	73
ATF2_47_59	P17544	51

Phosphonet

With the knowledge of peptides included in the chip, phosphonet file can be generated by Phosphonet scraper. Following columns are required in any order.

substrate	site	kinase_rank	kinase_name	kinase_id	kinexus_score_v2
P15336	69	1	JNK1 (MAPK8)	P45983	633
P15336	69	2	JNK3 (MAPK10)	P53779	633
P15336	69	3	ERK2 (MAPK1)	P28482	615
P15336	69	4	ERK1	P27361	611

Known Issues and limitations

• Script is only able to calculate kinase activity comparing two groups (control vs. experimental)
• Pamgene and Phosphonet files are not supplied, users are required to generate their own
• In current release the peptide kinetics matrix is automatically scaled and centered, if other beheviour is desired, it has to be changed in the code directly