fix Table S2 references

content/02.body.md

@@ -150,7 +150,7 @@ https://github.com/greenelab/xswap-analysis/blob/4f06bdaf1f034af9136e25c03f9891a
 
 We performed three prediction tasks to assess the performance of the edge prior.
 We compared the permutation-based prior with two additional predictors: our analytical approximation of the edge prior and the product of source and target degree, scaled to the range [0, 1] so that we could assess its calibration as well as its discrimination.
-We used 20 biomedical networks from the Hetionet heterogeneous network [@doi:10.7554/eLife.26726] that had at least 2000 edges for the first two tasks ([Table](#networks)).
+We used 20 biomedical networks from the Hetionet heterogeneous network [@doi:10.7554/eLife.26726] that had at least 2000 edges for the first two tasks (Table [S2](#tbl:networks)).
 
 In the first task, we computed the degree-based predictors (edge prior, scaled degree product, and analytical prior approximation) and predicted the original edges in the network by rank-ordering node pair edge predictions by the node pairs' predictor values.
 We used node pairs that lacked an edge in the original network as negative examples and those with an edge as positive examples.
@@ -232,7 +232,7 @@ https://github.com/greenelab/xswap-analysis/blob/4f06bdaf1f034af9136e25c03f9891a
 
 We evaluated degree as an edge prediction feature using the edge prior.
 In the first prediction task, we computed three predictors---the XSwap edge prior, an analytical approximation to the edge prior, and the (scaled) product of source and target node degree---on networks from Hetionet.
-We then evaluated the extent to which these predictors---treated as predictions themselves---could reconstruct the 20 networks (Table @tbl:xswap).
+We then evaluated the extent to which these predictors---treated as predictions themselves---could reconstruct the 20 networks (Table [S2](#tbl:networks)).
 The XSwap-derived edge prior reconstructed many of the networks with a high level of performance, as measured by the AUROC.
 Of the 20 individual networks we extracted from Hetionet, 17 had an edge prior self-reconstruction AUROC >= 0.95, with the highest reconstruction AUROC at 0.9971 (network was the Compound–downregulates–Gene edge type).
 Meanwhile, the lowest self-reconstruction performance (AUROC = 0.7697) occurred in the network having the fewest node pairs (network was the Disease–localizes–Anatomy edge type).

content/90.back-matter.md

@@ -195,7 +195,7 @@ Because the modified form of the approximation offers a much superior fit to the
 .tg .tg-s268{text-align:left}
 .tg .tg-0lax{text-align:left;vertical-align:top}
 </style>
-<table class="tg">
+<table id="tbl:networks" class="tg" data-tag="S2">
   <caption>
   <b>Table S2:</b>
   Networks used for the comparison.