COGs (Clusters of Orthologous Genomes) is a big database, which consists of thousands of alignments. Every alignment contains from 2 to almost infinite quantity sequences. Why there are no alignments from 1 sequence? To those who ask such questions, we strongly recommend to apply to the phylosophy of Multiple sequence alignments.
Why every alignment can be almost infinite? Well, every thing in this world is finite, though... It's very diffucult to look through big alignments, such as, for example, COG0495 (even though it is almost empty) - they contain too much lines. Imagine how you'll look through the giant, which unites 90 or 100 sequences. It will really look infinite!
That's why we consider that alignments contain from 2 to infinitely many sequences. Despite this, Clusters of Orthologous Genomes is a very suitable way of presenting multiple sequence alignments. An example of such alignment is shown on the same page COG0495.
This base - COGs - is permanently enlarging. New sequences are discovered, new alignments are made and added into COGs. More detailly about COGs one can know, for example, here.

COGs is a widely-used database. So, it's no use in searching new material for researcing - every new alignment is sure to appear in COGs database. Moreover, lots of different researches, based on COGs, are made. Thus, one can not only be content with this modest complex, but also join himself to the results of international scientific researches, compare them and make some interesting conclusions.