Zfisher is a web application to check the coding potentials of dubious genes in sequenced genomes based on the Z curve method. In the Zfisher, 33 features corresponding to single nucleotide and dinulceotide frequencies are used to describe the sequences. The Fisher linear discrimination algorithm is used to optimally differentiate protein-coding and non-coding sequences.
To run the tool, two types of sets are needed. First, DNA sequences of function-known genes would be uploaded or pasted into the upper window and they are used as training set of the method. Second, DNA sequences of doubious genes could be submitted to the lower window. Finally, the result will be shown in one new webpage several minutes after you clicking the Run button. In the result, five-fold validation and self-test based on the whole training set will be given, and details of the predicted non-coding ORFs are also provided.

Please upload the sequences of function-known genes in Fasta format
And upload the sequences of dubious hypothetical ORFs in Fasta format

Or

Please paste the sequences of function-known genes in Fasta format

And paste the sequences of dubious hypothetical ORFs in Fasta format

Appendix
Table 1. Details of the 413 ORFs eliminated from the RefSeq annotations in the ten complete genomes.
Table 2. Details of the 108 newly found genes in the ten Neisseriaceae strains.
Table 3. Details of assigned functions for each of the ten Neisseriaceae strains.
Table 4. Details of interrupted ORFs identified in seven Neisseriaceae strains.

Citation: Guo FB, Xiong L, Teng JL, Yuen KY, Lau SK*, Woo PC*. (2013) Re-annotation of protein-coding genes in 10 complete genomes of Neisseriaceae family by combining similarity-based and composition-based methods. DNA Res. 20(3): 273-86. [Full Text]