Factor VII is a member of a family of vitamin K-dependent, gamma-carboxylated plasma protein which includes factor IX, factor X, protein C, protein S and prothrombin. Activated factor VII (factor Vila) is a plasma serine protease which participates in a cascade of reactions leading to the coagulation of blood. Two overlapping genomic clones containing sequences encoding human factor VII were isolated and characterized. The complete sequence of the gene was determined and found to span 12.8 kilobases. The mRNA for factor VII as demonstrated by cDNA cloning is polyadenylated at multiple sites but contains only one AAUAAA poly-A signal sequence. The mRNA can undergo alternative splicing forming one transcript containing eight segments as exons and another with an additional exon which encodes a larger pre-pro leader sequence. The portion of the pre-pro leader coded for by the additional exon has no known counterpart in the other vitamin K-dependent proteins. The positions of the introns with respect to the amino acid sequence encoded by the eight essential exons of factor VII are the same as those present in factor IX, factor X, protein C and the first three exons of prothrombin. These exons code for domains generally conserved among members of this gene family, including a pre-pro leader (the essential exon la and alternative exon lb), a gamma-carboxylated domain (exons 2 and 3) a growth factor domain (exons 4 and 5) an activation region (exon 6) and a serine protease (exon 8). The corresponding introns in these genes are dissimilar with respect to size and sequence, with the exception of the third intron in factor VII and protein C. Four introns and a portion of exon 8 in factor VII contain regions made up of tandem repeats of oligonucleotide monomer elements. More than a quarter of the intron sequences and more than a third of the 3' untranslated portion of the mRNA transcript consist of these minisatellite tandem repeats. This type of structure is responsible for polymorphisms due to allelic variation in repeat copy number in other areas of the human genome. Tandem repeats can evolve as a result of random crossover in DNA whose sequence is not maintained by selection. This suggests that much of the sequence information present in the introns and untranslated portion of the message is dispensable.