DNA序列分析的生物信息学方法

5星 · 超过95%的资源需积分: 31 36 浏览量更新于2024-07-20 收藏 19.34MB PDF 举报

身份认证购VIP最低享 7 折!

30元优惠券

"《Bioinformatics for DNA Sequence Analysis》是由David Posada编辑的一本书，属于SpringerProtocols的Methods in Molecular Biology系列，专注于生物信息学在DNA序列分析中的应用。这本书详细介绍了利用生物信息学工具和R语言进行DNA序列分析的方法和技术。" 在生物科学领域，DNA序列分析是理解遗传信息、研究基因功能、揭示物种演化关系以及疾病发生机制的关键步骤。随着测序技术的飞速发展，我们如今能够处理和分析大量的DNA数据，这便催生了生物信息学这一跨学科领域的兴起。《Bioinformatics for DNA Sequence Analysis》一书深入探讨了如何利用生物信息学工具对DNA序列进行高效分析。书中可能涵盖了以下核心知识点： 1. **序列比对**：这是DNA分析的基础，包括全局比对（Global Alignment）和局部比对（Local Alignment），如Smith-Waterman算法和Needleman-Wunsch算法，用于寻找两个或多个序列间的相似性和差异性。 2. **进化模型**：书中可能讨论了不同的进化模型，如Jukes-Cantor模型、Kimura双参数模型和GTR模型，这些模型用于计算序列之间的进化距离。 3. **系统发生树构建**：通过最大简约法（Maximum Parsimony）、最大似然法（Maximum Likelihood）或贝叶斯法（Bayesian Inference）来重建物种间的进化关系。 4. **种群遗传学分析**：包括种群结构分析、遗传多样性测量和突变模式分析，帮助理解种群动态和遗传漂变。 5. **基因预测**：书中可能介绍了如何使用软件工具预测开放阅读框（Open Reading Frames, ORFs）和编码蛋白质的序列，以及非编码区域的功能。 6. **序列变异检测**：SNP（单核苷酸多态性）和indel（插入/缺失）的检测方法，这些变异在遗传疾病研究和个性化医疗中具有重要意义。 7. **R语言的应用**：R是一种强大的统计计算和图形展示语言，广泛用于生物信息学数据分析，书中会教授如何使用R进行数据预处理、统计分析和结果可视化。 8. **软件工具和数据库**：介绍常用的生物信息学软件（如BLAST、ClustalW、PHYLIP等）和数据库（如GenBank、NCBI、Ensembl等），并指导读者如何有效利用它们。 9. **案例研究和实战项目**：可能包含实际DNA序列分析的案例，让读者能够将所学知识应用于实际问题解决。 10. **最新技术与进展**：可能涵盖下一代测序技术（Next-Generation Sequencing, NGS）的数据处理和分析，以及云计算和大数据在生物信息学中的应用。此书为生物学家、遗传学家、计算机科学家以及对生物信息学感兴趣的学者提供了一本实用的指南，帮助他们掌握DNA序列分析的理论和实践技能。通过学习，读者不仅可以理解生物信息学的基本原理，还能熟练运用相关工具解决实际科研问题。

资源详情

资源推荐

2.1.2. Flat File Format GenBank, EMBL, and DDBJ each have their own flat file format,

but contain basically the same information. They are all based

upon the Feature Table, which can be found at http://

www.ncbi.nlm.nih.gov/collab/FT. For references to these file

types, see (6–9).

2.1.3. Accession Numbers,

Version Numbers, Locus

Names, Database

Identifiers, etc.

The standard for identifying a nucleotide sequence record is by an

accession.version system where the accession number is an identifier

of two letters followed by six digits and the version is an incre-

mental number indicating the number of changes that have been

Table 1.1

FASTA File sequence identifiers. Information from the NCBI

Handbook (25)

Database name Identifier syntax

GenBank gb|accession.version

EMBL emb|accession.version

DDBJ dbj|accession.version

NCBI RefSeq ref|accession.version

PDB pdb|entry|chain

Patents pat|country|number

NBRF PIR pir||entry

SWISS-PROT sp|accession|entry

Protein Research Foundation prf|name

GenInfo Backbone Id bbs|number

General database identifier gnl|database|identifier

Local Sequence identifier lcl|identifier

Table 1.2

IUB/IUPAC nucleotide and ambiguity codes

A adenosine M A or C (amino) V A, C, or G

C cytidine K G or T (keto) H A, C, or T

G guanine R A or G (purine) D A, G, or T

T thymidine Y C or T (pyrimidine) B C, G, or T

U uridine S A or T (strong) – Gap of indeterminate length

WCorG(weak) N A, C, G, or T (any or unknown)

4 Menlove, Clement, and Crandall

made to the sequence since it was first submitted. Locus names

(see Note 1) are older, less standardized identifiers whose original

purpose was to group entries with similar sequences (10). The

original locus format was intended to hold information about the

organism and other common group characteristics (such as gene

product). That ten-character format is no longer able to hold such

information for the large number and variety of sequences now

available, so the locus has become yet another unique identifier

often set to be the same value as the accession number. Database

identifiers are simply two- or three-character strings that serve to

indicate which database originally received and stored the infor-

mation. The database identifier is the first value listed in the

FASTA identifier syntax (Table 1.1).

When a sequence is first submitted to GenBank, it is submitted

with several defined features associated with the sequence. Some

include CDS (coding sequence), RBS (ribosome binding site),

rep_origin (origin of replication), and tRNA (mature transfer

RNA) information. A translation of protein coding nucleotide

sequences into amino acids is provided as part of the features

section. Likewise, labeling of different open reading frames,

introns, etc., are all part of the table of features. A list of features

and their descriptions, formats, and conventions that were agreed

upon by INSDC can be found in the Feature Table (see

Section 2.1.2).

2.2. Smith–Waterman

and Dynamic

Programming

In 1970, Needleman and Wunsch adapted the idea of dynamic

programming to the difficult problem of global sequence align-

ment (11). In 1981, Smith and Waterman adapted this algorithm

to local alignments (12). A global alignment attempts to align two

sequences throughout their entire length, whereas a local align-

ment aligns regions of two sequences where high similarity is

observed. Both methods involve initializing, scoring, and tracing

a matrix where the rows and columns correspond to the bases or

residues of the two sequences being aligned (Fig. 1.2). In the local

alignment case, the first row and the first column are filled with

zeroes. The remaining cells are filled with a metric value recursively

derived from neighboring values:

max

left neighbor þ gap penalty

top neighbor þ gap penatly

top-left neighbor þ match/mismatch score

If the current cell corresponds to a match (identical bases), the

match score is added to the value from the diagonal neighbor,

otherwise the mismatch score is used. The gap penalty and mis-

match scores are generally zero or a small, negative number while

the match score is a positive number, larger in magnitude. This

Similarity Searching Using BLAST 5

method is used recursively, starting from the upper left corner of the

matrix and proceeding to the lower right corner. Figure 1.2b and c

shows matrices from two different sets of gap and match scores.

To find a local alignment, one simply finds the largest number

in the matrix and traces a path back until a zero is reached, each

step moving to a cell that was responsible for the current cell’s

value. While this method is robust and is guaranteed to give the

best alignment(s) for a given set of scores and penalties, it is

important to note that often multiple paths and therefore multiple

alignments are possible for any given matrix when these para-

meters are used. As an example, b and c of Fig. 1.2 only differ

slightly in their gap and match scores, but produce very different

alignments. In addition, the set of scores and penalties used

dramatically affect the alignment, and finding the optimal set is

neither trivial nor deterministic. Weight matrices for protein-coding

sequences were developed in the late 1970s in an attempt to over-

come these challenges.

2.3. Weighting/Models

2.3.1. PAM Matrices

To increase the specificity of alignment algorithms and provide a

means to evaluate their statistical significance, it was necessary to

implement a meaningful scoring scheme for nucleotide and amino

acid substitutions. This was especially true when dealing with protein

(or protein-coding) sequences. In 1978, Dayhoff et al. developed

the first scoring or weighting matrices created from substitutions

that have been observed during evolutionary history (13).These

substitutions, since they have been allowed or accepted by natural

Fig. 1.2. Smith-Waterman local alignment example. (A) shows an empty matrix, initialized for a Smith-Waterman

alignment. (B) and (C) are alignments calculated using the specified scoring parameters. The alignment produced in

(B) is drastically different from that in (C), though they only differ slightly in their scoring parameters, one using a match

score of 1 and the other 2.

6 Menlove, Clement, and Crandall

selection, are called accepted point mutations (PAM). For

Dayhoff’sPAMmatrices,groupsofproteinswith85%ormore

sequence similarity were analyzed and their 1,571 substitutions

were cataloged. Each cell of a PAM matrix corresponds to the

frequency in substitutions per 100 residues between two given

amino acids. This frequency is referred to as one PAM unit. Back

in the 1970s, when they were created, however, there was a

limited number and variety of protein sequences available, so

theyarebiasedtowardsmall,globular proteins. It is also impor-

tant to note that each PAM matrix corresponds to a specific

evolutionary distance and that each is simply an extrapolation

of the original. For example, a PAM250 (Fig. 1.3)matrixis

constructed by multiplying the PAM1 matrix by itself 250

times and is viewed as a typical scoring matrix for proteins that

have been separated by 250 million years of evolution.

2.3.2. BLOSUM Matrices To overcome some of the drawbacks of PAM matrices, Henikoff

and Henikoff developed the BLOSUM matrices in 1992 (14).

These matrices were based on the BLOCKS database, which orga-

nizes proteins into blocks, where each block, defined by an align-

ment of motifs, corresponds to a family. Whereas the original PAM

matrix was calculated with proteins with at least 85% identity,

BLOSUM matrices are each calculated separately using conserved

motifs at or below a specific evolutionary distance. This diversity of

matrices coupled with being based on larger datasets makes the

BLOSUM matrices more robust at detecting similarity at greater

evolutionary distances and more accurate, in many cases, at per-

forming local similarity searches (15).

2.3.3. Choosing a Matrix When choosing a matrix, it is important to consider the alterna-

tives. Do not simply choose the default setting without some initial

consideration. In general, finding similarity at increasing diver-

gence corresponds to increasing PAM matrices (PAM1, PAM40,

PAM120, etc.) and decreasing BLOSUM matrices (BLOSUM90,

Fig. 1.3. PAM250 and BLOSUM45 substitution matrices.

Similarity Searching Using BLAST 7

BLOSUM80, BLOSUM62, etc.) (16). PAM matrices are strong

at detecting high similarity due to their use of evolutionary infor-

mation. However, as evolutionary distance increases, BLOSUM

matrices are more sensitive and accurate than their PAM counter-

parts. Table 1.3 includes a list of suggested uses.

2.4. BLAST Programs Nucleotide–nucleotide searches are beneficial because no infor-

mation is lost in the alignment. When a codon is translated from

nucleotides to amino acid, approximately 69% of the complexity

is lost (64 possible nucleotide combinations mapped to 20 amino

acids). In contrast, however, the true physical relationship

betweentwocodingsequencesisbestcapturedinthetranslated

view. Matrices that take into account physical properties, such as

PAM and BLOSUM, can be used to add power to the search.

Additionally, in a nucleotide search, there are only four possible

characterstatescomparedto20inanaminoacidsearch.Thusthe

probability of a match due to chance versus a match due to

common ancestry (identify in state versus identical by descent)

is high.

The Basic Local Alignment and Search Tools (BLAST) are the

most widely used and among the most accurate in detecting

sequence similarity (17)(see Note 2). The standard BLAST pro-

grams are Nucleotide BLAST (blastn), Protein BLAST (blastp),

blastx, tblastn, and tblastx. Others have also been developed to

meet specific needs. When choosing a BLAST program, it is

Table 1.3

Suggested uses for common substitution matrices. The matrices highlighted in bold

are available through NCBI’s BLAST web interface. BLOSUM62 has been shown to

provide the best results in BLAST searches overall due to its ability to detect large

ranges of similarity. Nevertheless, the other matrices have their strengths. For

example, if your goal is to only detect sequences of high similarity to infer homology

within a species, the PAM30, BLOSUM90, and PAM70 matrices would provide the

best results. This table was adapted from results obtained by David Wheeler (16)

Alignment size Best at detecting Similarity (%) PAM BLOSUM

Short Similarity within a species 75–90 PAM30 BLOSUM95

" Similarity within a genus 60–75 PAM70 BLOSUM85

Medium Similarity within a family 50–60 PAM120 BLOSUM80

" The largest range of similarity 40–50 PAM160 BLOSUM62

Long Similarity within a class 30–40 PAM250 BLOSUM45

" Similarity within the twilight zone 20–30 BLOSUM30

8 Menlove, Clement, and Crandall

剩余357页未读，继续阅读

weixin_38298121

粉丝: 0
资源: 1

DNA序列分析的生物信息学方法

Bioinformatics Data Skills Reproducible and Robust Research with Open 无水印pdf

2018基因检测行业研究报告-鲸准研究院-2018.06-56页.pdf

python for bioinformatics chapman & hall/crc pdf

pathbetweennodes MATLAB在哪个包里

matlab 處理不同fasta

fastqc在linux下安装运行

ubantu安装fastqc

fastqc的升级版本

linux centos7安装fastqc

蛋白磷酸化位点预测网站

essential bioinformatics pdf

附带FastQC操作代码和运行程序及软件

BMC BioInformatics期刊格式

commom subsequence

To search for alternate channels that may provide the conda package you're looking for, navigate to

matlab中生物信息学分析工具

Bioinformatics期刊

sklearn.manifold

最新资源