Skip to Content

Instrukcja korzystania z Biblioteki

Serwisy:

Ukryty Internet | Wyszukiwarki specjalistyczne tekstów i źródeł naukowych | Translatory online | Encyklopedie i słowniki online

Translator:

Kosmos
Astronomia Astrofizyka
Inne

Kultura
Sztuka dawna i współczesna, muzea i kolekcje

Metoda
Metodologia nauk, Matematyka, Filozofia, Miary i wagi, Pomiary

Materia
Substancje, reakcje, energia
Fizyka, chemia i inżynieria materiałowa

Człowiek
Antropologia kulturowa Socjologia Psychologia Zdrowie i medycyna

Wizje
Przewidywania Kosmologia Religie Ideologia Polityka

Ziemia
Geologia, geofizyka, geochemia, środowisko przyrodnicze

Życie
Biologia, biologia molekularna i genetyka

Cyberprzestrzeń
Technologia cyberprzestrzeni, cyberkultura, media i komunikacja

Działalność
Wiadomości | Gospodarka, biznes, zarządzanie, ekonomia

Technologie
Budownictwo, energetyka, transport, wytwarzanie, technologie informacyjne

Evaluation of Prediction Accuracy of Genefinders Using Mouse Genomic DNA

Six gene-finding programs i.e., Genscan, GeneMark hmm., HMMgene, GenView2, FGENESH and FGENESH+ were evaluated using 24 well defined mouse single- and multiexon genes to predict the structure of protein coding genes. Our analyses indicated that different methods often produce different and sometimes contradictory-results. In the nucleotide level, the highest correlation coefficient (0.87) and approximate correlation (0.86) values and also the lowest correlation coefficient (0.67) and approximate correlation (0.67) values were detected only for FGENESH+ and GenView 2 programs, respectively. Furthermore, at the exon level, similar results were obtained. In general, our results at either the nucleotide or exon levels showed that FGENESH+ (HMM plus sequence similarity programs), provide a level of improvement over the ab initio gene prediction methods such as Genview 2 and suggested that, probably, FGENESH+ and also Genscan can be more helpful than the others. Meanwhile, based on phylogenetic tree, all ab initio genefinders, excepted of GeneMarkhmm., were placed in the same group and FGENESH+ with GeneMarkhmm. programs assigned in another one. Moreover, based on our results, we realized that the accuracy of these programs, is strongly dependent on GC content. At last, on the basis of whole known sequences it was concluded that predictive accuracy of these programs is lower than actual.

Trends in Bioinformatics 2011/06/29 - 01:19 Czytaj