Build Clusters Based on PTM Data Under Like Conditions

This function takes the post-translational modification table and runs it through three calculations of distance: Euclidean Distance, Spearman Dissimilarity (1 - |Spearman Correlation|), and the average of the two of these. These calculations find the 'distance' between ptms based upon under what conditions they occur. These matricies are then run through t-SNE in order to put them into a 3-dimensional space. Additionally, the intersection of these 3 clusters is also created. A correlation table is also produced based on the Spearman Correlation table.

Usage

MakeClusterList(ptmtable, keeplength = 2, toolong = 3.5)

Arguments

ptmtable

A dataset for post-translational modifications. Formatted with row's names containing PTM names. The column names should be drugs. Values are numeric values that represent how much the PTM has reacted to the drug.

keeplength

Only keep clusters of ptms whose size is larger than this parameter. (I.e keeplength = 2 then keep ("AARS", "ARMS", "AGRS") but not ("AARS", "ARMS")); default is 2

toolong

A numeric threshold for cluster separation, defaults to 3.5.

Value

A list with these data structures at the given index:
1 (Consensus Clusters as a list): Clusters in all 3 distance metrices as a list.
2 (Consensus Clusters as an adjacent matrix) A matrix containing values of 0s and 1s depending on if the PTMs are cocluster with other PTMs, rows and columns are unamed.
3 (PTM Correlation Matrix) A data frame showing the correlation between ptms (as the rows and the columns) with NAs placed along the diagonal.

Details

Please note: t-SNE involves an element of randomness; in order to get the same results, set.seed(#) must be called.

Examples

Example_Output <- MakeClusterList(ex_tiny_ptm_table) #Run function
#> [1] "Starting correlation calculations and t-SNE."
#> [1] "2025-10-20 22:12:10 UTC"
#> [1] "This may take a few minutes for large data sets."



#> Starting FindCommonClusters at 2025-10-20 22:12:16.480787
#> [1] "2025-10-20 22:12:16 UTC"
#> [1] Total time: 0.0428409576416016
Example_Output[[1]][1:3] #Display data
#> $ConsensusCluster1
#> [1] "TNS3 p S602"  "CEBPB p S141"
#> 
#> $ConsensusCluster2
#> [1] "RACK1 p T229"   "ATP6V1H p Y388" "IRS2 p Y632"    "TNK2 p T829"   
#> 
#> $ConsensusCluster3
#> [1] "EPHA2 p Y588"   "ATG101 p Y164"  "SLC25A5 p Y195" "ALK p Y1584"   
#> 
#Do we want to have one for adj.consensus? Doesn't seem like it'd be very helpful to view.
utils::head(Example_Output[[3]][, c(1,2,3,4,5)]) #Display data
#>                           RNPS1 p Y205 PTPN6 p Y536 EIF2S1 p Y147
#> RNPS1 p Y205                        NA            1            NA
#> PTPN6 p Y536                         1           NA            NA
#> EIF2S1 p Y147                       NA           NA            NA
#> ACTG1 p T297; ACTB p T297           NA           NA            NA
#> RACK1 p T229                        NA           NA            NA
#> EML4 p Y453                         -1           -1            NA
#>                           ACTG1 p T297; ACTB p T297 RACK1 p T229
#> RNPS1 p Y205                                     NA           NA
#> PTPN6 p Y536                                     NA           NA
#> EIF2S1 p Y147                                    NA           NA
#> ACTG1 p T297; ACTB p T297                        NA           NA
#> RACK1 p T229                                     NA           NA
#> EML4 p Y453                                      NA           NA