Mahmood Akhtar

Signal Processing in Sequence Analysis: Advances in Eukaryotic Gene Prediction

Genomic sequence processing has been an active area of research for the past two decades and has increasingly attracted the attention of digital signal processing researchers in recent years. A challenging open problem in deoxyribonucleic acid (DNA) sequence analysis is maximizing the prediction accuracy of eukaryotic gene locations and thereby protein coding regions. In this paper, DNA symbolic-to-numeric representations are presented and compared with existing techniques in terms of relative accuracy for the gene and exon prediction problem. Novel signal processing-based gene and exon prediction methods are then evaluated together with existing approaches at a nucleotide level using the Burset/Guigo1996, HMR195, and GENSCAN standard genomic datasets. A new technique for the recognition of acceptor splice sites is then proposed, which combines signal processing-based gene and exon prediction methods with an existing data-driven statistical method. By comparison with the acceptor splice site detection method used in the gene-finding program GENSCAN, the proposed DSP-statistical hybrid technique reveals a consistent reduction in false positives at different levels of sensitivity, averaging a 43% reduction when evaluated on the GENSCAN test set.

On DNA Numerical Representations for Period-3 Based Exon Prediction

Processing of DNA sequences using traditional digital signal processing methods requires their conversion from a character string into numerical sequences as a first step. Many representations introduced previously assign values to indicate the four DNA nucleotides A, C, G, and T that impose mathematical structures not present in the actual DNA sequence. In this paper, almost all existing methods are compared for the purpose of identifying protein coding regions, using the discrete Fourier transform (DFT) based spectral content measure to exploit period-3 behaviour in the exonic regions for the GENSCAN test set. False positive vs. sensitivity, receiver operating characteristic (ROC) curve and exonic nucleotides detected as false positive results all show that the two newly proposed numerical of DNA representations perform better than the well-known Z-curve, tetrahedron, and Voss representations, with 66-75% less processing. By comparison with Voss representation, the proposed paired numeric method can produce relative improvements of up to 12% in terms of prediction accuracy of exonic nucleotides at a 10% false positive rate using the GENSCAN test set.

Optimizing period-3 methods for eukaryotic gene prediction

In this paper, we firstly investigate the effect of window lengths on selected signal processing-based gene and exon prediction methods. We then optimize these methods to improve their prediction accuracy by employing the best DNA representation, a suitable window length, and boosting the output signals to enhance protein coding and suppress the non-coding regions. It is shown herein that the proposed method outperforms major existing time-domain, frequency- domain, and combined time-frequency approaches. By comparison with the existing DFT-based methods, the proposed method not only requires 50% less processing but also exhibits relative improvements of 53.3%, 46.7%, and 24.2% respectively over spectral content, spectral rotation, and paired and weighted spectral rotation measures in terms of prediction accuracy of exonic nucleotides at a 5% false positive rate using the GENSCAN test set.

Gene and exon prediction using time domain algorithms

The protein-coding regions of DNA sequences normally exhibit a period-3 behaviour that can be used to identify gene locations. Various methods have been used to automatically identify the coding regions, however these have been predominantly ‘frequency’ domain techniques. Numerous ‘time’ domain techniques are available from the signal processing literature, and this work examines their applicability to gene and exon prediction in DNA sequences. Two techniques new to this application are introduced: the Time Domain Periodogram (TDP) and the Average Magnitude Difference Function (AMDF). We also present an indicative comparison of time domain and existing frequency domain techniques, from which the AMDF appears to be the most promising technique. Autoregressive modelling methods are further investigated in this work.

An integer period DFT for biological sequence processing

Detection of periodicity in symbolic sequences such as DNA is of considerable interest in a number of applications, however fast, accurate algorithms are needed for measuring spectral content at multiple integer periods. This paper describes an integer period discrete Fourier transform (IPDFT), presents a new algorithm for its implementation, and discusses applications to DNA sequence analysis. Evaluations on DNA sequence data show that the IPDFT may be a more suitable tool for periodicity analysis than an existing widely used correlation-based approach.

A motion adaptive animal chamber for PET imaging of freely moving animals

Small animal positron emission tomography (PET) is a potentially powerful tool for understanding the molecular origins of debilitating brain disease such as dementia, depression and schizophrenia. However, its full potential in such investigations has not yet been realized due to the use of anaesthesia to avoid motion artifacts. Anaesthesia alters biochemical pathways within the brain and precludes the study of animal behavior during the imaging study. Previously we have reported a motion correction approach for conscious animal PET imaging that employs motion tracking and line of response (LOR) rebinning. We are currently extending this technique to allow PET imaging of freely moving animals, enabling the non-invasive measurement of biochemical processes in the brain of a fully conscious rat while simultaneously observing its behavior. As a first step we report a robot-controlled motion adaptive animal chamber which translates in the horizontal plane based on the head position reported by a motion tracking system to compensate for gross animal movement and keep the head within the field of view (FOV) as long as possible during the scan. In a pilot animal study within a simulated microPET environment, the control algorithm increased the time the head spent centrally in the FOV from 38% to 83% without any apparent disturbance to the animals behaviour. We conclude that a robot-controlled motion adaptive chamber is a feasible approach and an important step towards imaging freely moving animals.

Time and Frequency Domain Methods for Gene and Exon Prediction in Eukaryotes

The detection of period-3 components in exons of eukaryotic gene sequences enables signal processing based time-domain and frequency-domain methods to predict these regions. In this paper, we improve the prediction accuracy of frequency-domain methods by proposing a new algorithm known as the paired and weighted spectral rotation (PWSR) measure, which exploits both period-3 behaviour and another useful statistical property of genomic sequences. By comparison with existing frequency-domain approaches, the proposed PWSR method reveals relative improvements of 15.2% and 10.7% respectively over spectral content and spectral rotation measures in terms of prediction accuracy of exonic nucleotides at a 10% false positive rate using the GENSCAN test set. Finally, we combine the proposed PWSR with an existing time-domain method to demonstrate further signal processing-based improvements in gene and exon prediction accuracy.

Detection of period-3 behavior in genomic sequences using singular value decomposition

Manydigital signal processing techniques have beenusedtoautomatically distinguish theprotein coding regions (exons) fromnon-coding regions (introns) inaDNAsequence. Recently, Auto-regressive (AR) technique hasbeenused forthedetection of3periodicty present inprotein coding regions of genomic sequences. The. sequence length, average spacing between coding regions, andaverage coding region length arethemainfactors thataffect the performance ofanyprediction method Inthis paper, wepropose theuseofSingular Value Decomposition (SVD) methodfor thedetection ofperiod-3 behavior in DNA sequences. Results showthatSVDmethod outperforms theARtechnique.

Motion tracking of fully conscious small animals in PET

Pre-clinical positron emission tomography (PET) is becoming increasingly important in understanding brain physiology using animal models. One of the major challenges at present is being able to perform brain PET studies without the use of anesthesia. In most cases where the animal is minimally restrained this will require some form of motion tracking to provide the necessary temporal pose information for compensation before or during image reconstruction (eg. [1, 2]). In previous work we have demonstrated successful tracking of continuous movement in phantom studies and an anesthetized rat study in which the animal was moved manually [2]. Here we report on our first trials tracking the head of fully conscious rats moving continuously during emission and transmission PET acquisitions performed on a microPET Focus 220 scanner (Siemens Preclinical Solutions, Knoxville, USA). The motion tracking is based on a commercial stereo-optical motion tracking device called the MicronTracker Sx60 (Claron Tech. Inc., Toronto, Canada). We have previously reported in detail on this device and its suitability for small-scale motion tracking [3]. In this paper we (i) describe the motion tracking setup and marker considerations used for conscious animal head tracking, (ii) describe our animal handling methods, (iii) present data on motion tracker performance in the conscious animal trials, (iv) present data and observations on the character of rat movements in the imaging environment, and (v) show examples of the correction that can be obtained using these data for motion compensation.

Digital Signal Processing Techniques for Gene Finding in Eukaryotes

In this paper, we investigate the effects of window shape and length on a DFT-based method for gene and exon prediction in eukaryotes. We then propose a new gene finding method which combines the selected time-domain and frequency-domain methods, by employing the most effective DNA symbolic-to-numeric representation examined to date in conjunction with suitable window shape and length parameters and a signal boosting technique. It is shown herein that the new method outperforms major existing approaches. By comparison with the existing methods, the proposed method reveals relative improvements of 15.1% to 55.9% over different methods in terms of prediction accuracy of exonic nucleotides at a 5% false positive rate using the GENSCAN test set.