---
name: "eval_predictions"
section: "1"
platform: "linux"
package: "phast"
version: "1.7+dfsg-2"
release: "Debian 13 (Trixie)"
platforms: ["linux"]
url: "https://new.carta.tech/eval_predictions"
---

# eval\_predictions(1)

Compares predicted genes with "real" (or annotated) genes.

## NAME

eval\_predictions - Compares predicted genes with "real" (or annotated) genes.

## SYNOPSIS

**eval\_predictions** **-r** \<real\_fname\_list\> **-p** \<pred\_fname\_list\> **-l** \<seq\_len\_list\> \[OPTIONS\]

## DESCRIPTION:

Compares predicted genes with "real" (or annotated) genes. Reports standard measures of prediction quality.  The following measures are reported:

-

  \- nucleotide sensitivity (Sn)

-

  \- nucleotide specificity (Sp)

-

  \- approximate correlation (AC)

-

  \- correlation coefficient (CC)

-

  \- exon sensitivity (ESn)

-

  \- exon specificity (ESp)

-

  \- proportion of real exons correctly predicted (CRa)

-

  \- proportion of real exons partially predicted (PCa)

-

  \- proportion of real exons with overlapping predictions (OLa)

-

  \- missed exons (ME)

-

  \- proportion of predicted exons that are correct (CRp)

-

  \- proportion of predicted exons that are partially correct (PCp)

-

  \- proportion of predicted exons that overlap real ones (OLp)

-

  \- wrong exons (WE)

All quantities are computed as described in "Evaluation of Gene-Finding Programs on Mammalian Sequences," by Rogic et al. (Genome Research 11:817-832).  Note that CRa + PCa + OLa + ME = 1 and CRp + PCp + OLp + WE = 1.  Note also that each set (predicted and real) should consist of non-overlapping groups of features (see 'refeature').

## OPTIONS

**-r** \<real\_fname\_list\>

-

  (required) List of names of files defining real genes (GFF).

**-p** \<pred\_fname\_list\>

-

  (required) List of names of files defining predicted genes (GFF).  Must correspond in order to \<real\_fname\_list\>.

**-l** \<seq\_len\_list\> (required) List of lengths of sequences. Needed to compute

-

  certain nucleotide-level statistics.

**-f** \<feat\_list\>

- List of names of all features denoting exon regions.

  By

-

  default, equal to the single name "CDS".

**-d** \<fname\_prefix\> Dump full coords of correct, partially correct, wrong, missed, and overlapping exons to a set of files having the specified file name prefix.

**-n** \<nbases\> Also report stats on "nearly correct" exons, that is, incorrect exons whose boundaries are within \<nbases\> of being correct. Columns will be labeled "NCa" and "NCp".

**-h** Print this help message.

NOTE: be sure stop codons are included in CDSs in both the predicted and real sets, or in neither set.