Workflow used in this report. (a) AI text mining procedure was done in various phases. First, answers for the imposed question were collected. Secondly, answers were filtered and gene names were extracted. Thirdly, the resulting genes were clustered according to their reported function into a paracellular pathway component, transcellular component, or both. Fourthly, phylogenetic analysis was done and lastly, functional conservation for critical genes was performed. (b) The genes resulting from the AI step were categorized based on literature review into one of three groups; shared genes (in navy blue) and paracellular (P) and transcellular (T) (in light red).

Paracellular versus transcellular migration pathways. Several genes are used exclusively by paracellular diapedesis route such as CX43, Claudin, Occludin, JAMs and VE-cadherin. Similarly, CAVs and VAMPs seem to be activated only in transcellular pathway. Genes that are activated in both pathways include those who function in transmembrane movements? such as RhoA and ?-catenin pathway.

Comparison between the phylogenetic history of the active components of paracellular and transcellular migration pathways. We investigated the phylogenetic of (a) PECAM1 the main active component of paracellular migration and (b) CAV1 the main active component of the transcellular migration pathway. PECAM1 diverged during bony fish emergence where CAV1 diverged during placozoa first appearance.

Comparative functional conservation between PECAM1 and CAV1. Pecam1 and CAV1, have low similarity. Symbols used are; Humans (H), Chimp (C), HM (house mouse), R (Red Jungle fowl), ZF (Zebrafish), ES (Elephant shark), (RW) Roundworm, Starlet sea anemone (SAA), SL (Sea lampreys) and TRX(Trichoplax).

Evolutionary network of PECAM1. We used HHserach, Blastp, and Splittrees to build an evolutionary network for PECAM1. We detected two putative ancestors for PECAM1 (VCAM1 and FCRL5). However, FCRL5 has the shortest evolutionary pathway to PECAM1.

Phylogenetic history of the main components of paracellular and transcellular migration. Dark blue represents homologs found in respective genes. Light blue represents orthologs found for a gene family. Light red represents non existing orthologs or homologs. Almost all exclusive transcellular route genes identified are ancient except for CAVLIN1 and PLVAP, which both first appear in Danio rerio. Most genes that are exclusively active in the paracellular route have recently diverged such as CX43, JAMs, JAML, Occludin, PECAM1, VE-cadherin. Genes that are active in both routes are mostly ancient such as RhoA, RhoG, SRC, SRF VASP, CDC42, CTNNB1 and VWF.