Since Alec Jeffreys coined the term "individual-specific genetic fingerprints" in 1985, DNA typing has become indispensable in forensic analysis, having as its central rationale the evidential power of the "match" between trace evidence and suspect. Established technological advances in this field include the adoption of the polymerase chain reaction (PCR) to generate DNA profiles from minute biological samples, the use of mitochondrial DNA to obtain information from old bones and the application of Y chromosomal polymorphisms in sexual assault cases.Present research questions focus on genetic markers for external visible traits and on improving the utility of poor quality samples with degraded DNA.The first valid marker for a physical trait was the sex marker amelogenin, which has been part of short tandem repeat (STR) multiplex PCRs for over a decade. Current studies are investigating markers for human pigmentation, predicting hair, iris and skin colour. The relationship between genome and pigmentation however is complex, involving numerous genes. Single nucleotide polymorphisms (SNPs) in these genes make promising candidates for predictive markers.Physical stature is also a complex genetic trait where study subjects can be phenotyped easily. Linkage analysis has revealed several short stature-quantitative trait loci, and polymorphisms in the fibrillin I gene are associated with tall stature.An indirect approach of physical traits is inferring the geographic origin of an individual from DNA. Here, genome-wide SNP panels have been used, correctly discriminating "European", "East Asian" and "Sub-Sahara African" ancestry, or even substructure amongst Europeans almost to the level of nationality, by typing 500 000 SNPs.The ultimate goal of this work is to achieve "molecular photofitting", using molecular techniques to construct a portrait of the person leaving a biological trace.The other direction of contemporary forensic DNA research addresses the difficulty of obtaining a DNA profile in a degraded crime stain.One strategy to tackle this problem uses SNPs in stead of STRs. Due to the intrinsically small size of SNPs, they are well suited for degraded samples, but there are two major drawbacks. First, due to their bi-allelic nature, the resolution of a SNP is inferior to any STR. This can be overcome by using larger numbers of SNPs, yielding the same statistical power as 10 to15 STRs. Then, since DNA profiles in national forensic databases currently contain only a standard set of STRs, a DNA profile consisting of SNPs alone cannot be matched.An alternative strategy would still apply STRs, but with smaller amplicons. Primer pairs are chosen closer to the repeat region, yielding shorter PCR products than with classical primers. The STRs in question are both "old", internationally used loci and newly developed systems. Thus, comparison with DNA profiles in existing forensic databases is possible.