Any gene or protein could have shared a common ancestor with any other gene or protein at some time in the past. But usually when we speak of homologs (othologs and paralogs) we have solid evidence of shared common ancestry because the sequences are far more similar to each other than could have happened by chance. There are hundreds of types of proteins for example that contain a zinc finger domain for binding to nucleic acid. This domain may or may not have evolved many time independently. Likewise there are dozens of homo box domain proteins which bind DNA to control development in multicellular organisms and these surely did share a common ancestor less than one billion years ago.

I am attaching screen shots of alignments of homobox proteins from eukaryotes and a selection of CAAD and PqqC proteins. If CAAD did share any ancestor with PccQ it would have been back in the primordial soup days.

The screen shots are from AliView, which is not a multiple sequence alignment program, but a multiple sequence alignment viewer Editor. AliView is free software for Mac OS-X. The equivalient program for Windows is BioEdit.

http://www.ormbunkar.se/aliview/

http://www.mbio.ncsu.edu/BioEdit/bioedit.html

Most phylogenetic analysis packages such as MEGA and DAMBE also have multiple sequence alignment viewers built in, but these stand alone programs have many more useful features.

Neither BLAST scores nor viewing an alignment can really tell us much about homology, but BLAST scores can be quite misleading because they are based both on sequence length and identity/similarity.

Both BioEdit and AliView have some multiple sequence alignment programs such as CLUSTAL built into them, but there are several other more powerful or better multiple sequence alignment programs available too, such as MAFFT and MUSCLE .

https://mafft.cbrc.jp/alignment/software/

https://www.drive5.com/muscle/