Cell spatial location differs between sdimx/y and CenterX/Y_global_px

[rbutleriii]

Hello, looking at the spatial locations themselves, and the values in the cell metadata, and I noticed the locations are close, but not exactly the same. This is a CosMx dataset built from their source files. Do you know where CenterX/Y_global_px is calculated in the process? (gobj@params at bottom)

> head(spat)
       sdimx     sdimy            cell_ID
       <num>     <num>             <char>
1: -158086.2 -32159.23  s23-fov021-cell_1
2: -158697.3 -32231.50  s23-fov021-cell_5
3: -158784.4 -32243.11  s23-fov021-cell_6
4: -158549.2 -32259.82  s23-fov021-cell_7
5: -158648.6 -32259.72  s23-fov021-cell_9
6: -158846.6 -32303.56 s23-fov021-cell_10
> head(meta[, 1:9])
              cell_ID list_ID   fov  Area AspectRatio CenterX_local_px CenterY_local_px CenterX_global_px CenterY_global_px
               <char>  <char> <int> <int>       <num>            <int>            <int>             <num>             <num>
1:  s23-fov021-cell_1     s23    21  8190        0.74             3105             3582         -158091.5         -32160.08
2:  s23-fov021-cell_5     s23    21  8691        1.07             2505             3502         -158691.5         -32240.08
3:  s23-fov021-cell_6     s23    21  8076        1.11             2404             3495         -158792.5         -32247.08
4:  s23-fov021-cell_7     s23    21 13561        1.22             2652             3475         -158544.5         -32267.08
5:  s23-fov021-cell_9     s23    21  3111        1.17             2544             3485         -158652.5         -32257.08
6: s23-fov021-cell_10     s23    21 12394        1.19             2355             3443         -158841.5         -32299.08
> head(spat[cell_ID %like% "s24"])
      sdimx     sdimy           cell_ID
      <num>     <num>            <char>
1: 162178.4 -16164.43 s24-fov034-cell_2
2: 161012.9 -16171.01 s24-fov034-cell_4
3: 162061.4 -16201.36 s24-fov034-cell_5
4: 161602.3 -16194.60 s24-fov034-cell_7
5: 162454.0 -16192.91 s24-fov034-cell_8
6: 161909.4 -16202.06 s24-fov034-cell_9
> head(meta[cell_ID %like% "s24", 1:9])
             cell_ID list_ID   fov  Area AspectRatio CenterX_local_px CenterY_local_px CenterX_global_px CenterY_global_px
              <char>  <char> <int> <int>       <num>            <int>            <int>             <num>             <num>
1: s24-fov034-cell_2     s24    34  9961        1.47             3732             4210          157921.8         -16164.39
2: s24-fov034-cell_4     s24    34  4181        1.11             2564             4204          156753.8         -16170.39
3: s24-fov034-cell_5     s24    34  6954        1.09             3618             4174          157807.8         -16200.39
4: s24-fov034-cell_7     s24    34  5498        1.38             3154             4180          157343.8         -16194.39
5: s24-fov034-cell_8     s24    34  3899        1.09             4007             4180          158196.8         -16194.39
6: s24-fov034-cell_9     s24    34  4594        1.06             3462             4172          157651.8         -16202.39
> head(spat[cell_ID %like% "s25"])
      sdimx     sdimy           cell_ID
      <num>     <num>            <char>
1: 446905.5 -59695.01 s25-fov017-cell_1
2: 447137.1 -59729.96 s25-fov017-cell_2
3: 446814.2 -59742.80 s25-fov017-cell_4
4: 447243.6 -59798.64 s25-fov017-cell_5
5: 446866.3 -59815.23 s25-fov017-cell_6
6: 446529.7 -59839.04 s25-fov017-cell_7
> head(meta[cell_ID %like% "s25", 1:9])
             cell_ID list_ID   fov  Area AspectRatio CenterX_local_px CenterY_local_px CenterX_global_px CenterY_global_px
              <char>  <char> <int> <int>       <num>            <int>            <int>             <num>             <num>
1: s25-fov017-cell_1     s25    17 10037        2.63             3472             4220          438392.8         -59701.88
2: s25-fov017-cell_2     s25    17 14817        1.24             3702             4190          438622.8         -59731.88
3: s25-fov017-cell_4     s25    17  6619        1.13             3385             4180          438305.8         -59741.88
4: s25-fov017-cell_5     s25    17 15248        0.73             3806             4124          438726.8         -59797.88
5: s25-fov017-cell_6     s25    17  8311        1.20             3432             4110          438352.8         -59811.88
6: s25-fov017-cell_7     s25    17  9894        0.89             3098             4079          438018.8         -59842.88

The final output is several horizontally joined gobjs, that were subset by region using the graphical selection tool, but I assume the global params would be unaltered after the join. Is the sdimx/y in the spat Locations more reliable?

gobj@parameters
$`0_subset`
            gobject           spat_unit           feat_type            cell_ids            feat_ids           poly_info      spat_unit_fsub 
          "gobject"         "spat_unit"         "feat_type" "selected_cell_ids" "selected_feat_ids"         "poly_info"    "spat_unit_fsub" 
     feat_type_ssub             verbose     toplevel_params 
   "feat_type_ssub"           "verbose"                 "2" 

$`1_filter`
               gobject              feat_type      expression_values   expression_threshold  feat_det_in_min_cells min_det_feats_per_cell 
                   "."                  "rna"                  "raw"                    "1"                  "100"                   "20" 
        spat_unit_fsub         feat_type_ssub              tag_cells          tag_cell_name              tag_feats         tag_feats_name 
               ":all:"                ":all:"                "FALSE"                  "tag"                "FALSE"                  "tag" 
               verbose 
                "TRUE" 

$`2_normalize`
          gobject         feat_type expression_values      norm_methods library_size_norm       scalefactor          log_norm 
              "."             "rna"             "raw"        "standard"            "TRUE"            "6000"            "TRUE" 
       log_offset           logbase       scale_feats       scale_cells       scale_order             theta       update_slot 
              "1"               "2"            "TRUE"            "TRUE"     "first_feats"             "100"          "scaled" 
          verbose 
           "TRUE" 

$`3_normalize`
          gobject         feat_type expression_values      norm_methods library_size_norm       scalefactor          log_norm 
              "."             "rna"             "raw"   "pearson_resid"            "TRUE"            "5000"            "TRUE" 
       log_offset           logbase       scale_feats       scale_cells       scale_order             theta       update_slot 
              "1"               "2"            "TRUE"            "TRUE"     "first_feats"             "100"         "pearson" 
          verbose 
           "TRUE" 

$`4_feat_stats`
            gobject           feat_type   expression_values detection_threshold      return_gobject 
                "."               "rna"        "normalized"                 "0"              "TRUE" 

$`5_cell_stats`
            gobject           feat_type   expression_values detection_threshold      return_gobject 
                "."               "rna"        "normalized"                 "0"              "TRUE" 

$`6_subset`
        gobject       spat_unit       feat_type        cell_ids  spat_unit_fsub  feat_type_ssub         verbose toplevel_params 
         "gobj"         ":all:"           "rna"         "cells"         ":all:"         ":all:"         "FALSE"             "2" 

$`7_subset`
        gobject       spat_unit       feat_type        cell_ids  spat_unit_fsub  feat_type_ssub         verbose toplevel_params 
         "gobj"         ":all:"           "rna"      "na_cells"         ":all:"         ":all:"         "FALSE"             "2" 

$`8_delaunay_spatial_network`
              dimensions used                        method    maximum distance threshold       name of spatial network 
"dimensions: sdimx and sdimy"                      "deldir"                        "auto"            "Delaunay_network"