Table 2.

Software function comparison in terms of data processing ability.

Software	Output Formats	Input Formats	Data Fault Tolerant a
UltraPse	SVM b, TSV c, CSV d	Multi-line FASTA (Automatic ID recognition for UniProt, GenBank, EMBL, DDBJ and RefSeq)	User-controllable behavior on data faults
PseAAC-General [46]	SVM, TSV, CSV	Single-line FASTA (With restrictions on comment line) e	Automatically ignore and report data faults
PseAAC-Builder [45]	SVM, TSV, CSV	Single-line FASTA (With restrictions on comment line)	Automatically ignore and report data faults
Pse-In-One [52]	SVM, TSV, CSV	Mutlti-line FASTA	Abort processing on data faults
PseKNC [48]	SVM, TSV, CSV	Mutlti-line FASTA	Abort processing on data faults
PseKNC-General [49]	SVM, TSV, CSV	Mutlti-line FASTA	Abort processing on data faults

^a Data fault tolerant: The behavior of a software when it encounters some invalid data records. Here, the invalid data records include the sequences with non-standard letter and the sequence without sufficient length; ^b SVM: data format for libSVM [61]; ^c TSV: tab separated vector; ^d CSV: comma separated vector; ^e Single-line FASTA: the sequence of a record in the file must not spread to multiple lines. Both PseAAC-General and PseAAC-Builder have the same restrictions.