Secuenciación de ADN mediante un algoritmo basado en hidden Markov models para el dispositivo MinION

Abstract

MinION devices are relatively modern sequencers that are characterized by their abysmally low price and dimensions compared to previous sequencers. For these reasons, the device is expected with great enthusiasm by biology researchers. However, the MinION has the big disadvantage

that the nanopore detection method it uses highly performs error prone DNA sequencing, which makes its use not widespread. Nevertheless, sequencing precision can be improved using base calling algorithms. In this work we develop a base calling algorithm based on a Hidden Markov Model as an approximation of a sequencing algorithm that improves the precision achieved so far. We check that this algorithm works properly for MinION devices that behave under our assumptions and we analyze the behavior of its precision curve based on the error of the device under ideal conditions. Moreover, we find that for measurements made with a real MinION device, very low sequencing accuracies are obtained with our base calling algorithm in relation to those expected. These results lead us to conclude that the actual device behaves in a non ideal way and more realistic assumptions are needed to develop a base calling algorithm for the actual device.