Version 3.7 (R2011a) Bioinformatics Toolbox Software

This table summarizes what's new in Version 3.7 (R2011a):

New Features and Changes	Version Compatibility Considerations	Fixed Bugs and Known Problems
Yes Details below	Yes—Details labeled as Compatibility Considerations, below. See also Summary.	Bug Reports Includes fixes

New and updated features in this version include:

• Data Format and Database Functions

• Sequence Statistics Functions

• Updates to the BioIndexedFile Class, Properties, and Methods

• Updates to BioRead and BioMap Classes and Methods

• Demos for High-Throughput Sequence Analysis

• Support Vector Machine (SVM) Functions

• Function Elements Being Removed

Data Format and Database Functions

The following functions have a new field, FilePath, in their output structure:

• fastainfo — Return information about FASTA file.

• fastqinfo — Return information about FASTQ file.

• saminfo — Return information about Sequence Alignment/Map (SAM) file.

The fastainfo function has two additional fields in its output structure: Header and Length.

Sequence Statistics Functions

Compatibility Consideration

In Bioinformatics Toolbox Version 3.6, the aacount and basecount functions still allowed 'Others' and 'Structure' name-value pairs, but displayed a warning.

In Bioinformatics Toolbox Version 3.7, the aacount and basecount functions do not allow 'Others' and 'Structure' name-value pairs, and return an error if you use them. Now you must use the 'Ambiguous' and 'Gaps' name-value pairs, which specify whether to count or ignore ambiguous characters and gaps, as well as specify how to count ambiguous characters, and whether to display a warning.

Updates to the BioIndexedFile Class, Properties, and Methods

The following name-value pairs of the BioIndexedFile constructor function are renamed:

• MapKeys is now IndexedByKeys.

• MemMapIndex is now MemoryMappedIndex.

Note   The former name-value pairs are still valid for Bioinformatics Toolbox Version 3.7 (R2011a).

The MemoryMappedIndex property of the BioIndexedFile class is now editable, which lets you load and unload file indices in memory.

The BioIndexedFile class includes the following new methods:

• getDictionary — Retrieve reference sequence names from SAM-formatted source file associated with BioIndexedFile object.

• getSubset — Create object containing subset of elements from BioIndexedFile object.

Updates to BioRead and BioMap Classes and Methods

The BioMap constructor includes a new name-value pair, SubsetRef, which lets you specify one reference sequence in the input argument (BioIndexedFile object, SAM-formatted file, or structure) when constructing the BioMap object.

The following method of the BioRead and BioMap classes is updated:

getSubset — Create object containing subset of elements from object. Updated with addition of the Indexed name-value pair, which lets you use the BioIndexedFile object when creating a new object, thus saving memory. This name-value pair is ignored if your BioRead or BioMap object was not created from a BioIndexedFile object.

Following are new methods of the BioMap class:

• getBaseCoverage — Return base-by-base alignment coverage of reference sequence in BioMap object.

• getCounts — Return count of read sequences aligned to reference sequence in BioMap object.

• getIndex — Return indices of read sequences aligned to reference sequence in BioMap object.

The getCoverage method of the BioMap class is being removed in a future release. Use the getBaseCoverage, getCounts, and getIndex methods instead.

Compatibility Considerations

In Bioinformatics Toolbox Version 3.6 and earlier, the BioMap class included a getCoverage method, which computes read coverage in a BioMap object.

In Bioinformatics Toolbox Version 3.7, the getCoverage method still runs, but displays a warning. Now use the getBaseCoverage, getCounts, and getIndex methods of the BioMap class.

Demos for High-Throughput Sequence Analysis

Following are two new high-throughput sequence analysis demos:

• Exploring Protein-DNA Binding Sites from Paired-End ChIP-Seq Data

• Identifying Differentially Expressed Genes from RNA-Seq Data

Support Vector Machine (SVM) Functions

The functionality of the svmsmoset function is incorporated into the svmtrain and statset functions. Although svmsmoset is still valid, it is no longer documented.

The svmtrain function has been updated:

• The function can now handle NaN values in the training matrix input and performs more checks of parameters you supply.

• The function now includes Sequential Minimal Optimization (SMO) functionality plus four new name-value pairs: kernelcachelimit, kktviolationlevel, options, and tolkkt.

• The default training method is SMO, even if you have Optimization Toolbox™ installed.

• The QuadProg_Opts and SMO_Opts name-value pairs have been replaced by the options name-value pair. Although the former name-value pairs are still valid, the recommended ways to perform quadratic programming (QP) training and SMO training are summarized in the following bullets.

• The recommended way to include QP options for svmtrain is to use the QP training method and use the new options name-value pair. For the options value, use a structure you create with optimset.

• The recommended way to include SMO options for svmtrain is to use the default SMO training method and use the new kernelcachelimit, kktviolationlevel, options, and tolkkt name-value pairs. For the options value, use a structure you create with the statset function and its Display and MaxIter name-value pairs.

Compatibility Considerations

In Bioinformatics Toolbox Version 3.6 and earlier, if you had Optimization Toolbox installed, QP was the default training method for the svmtrain function. Now the default training method is SMO.

Function Elements Being Removed

Function Element Name	What Happens When You Use This Function Element	Use This Instead	Compatibility Considerations
'Others' name-value pair as input to aacount and basecount functions	Errors	'Ambiguous' or 'Gaps' name-value pair as input to aacount and basecount functions	See the Compatibility Considerations subheading in Sequence Statistics Functions.
'Structure' name-value pair as input to aacount and basecount functions	Errors	'Ambiguous' name-value pair with either 'ignore' or 'warn' as input to aacount and basecount functions	See the Compatibility Considerations subheading in Sequence Statistics Functions.
getCoverage method of BioMap class	Warns	getBaseCoverage, getCounts, and getIndex methods	See the Compatibility Considerations subheading in Updates to BioRead and BioMap Classes and Methods.
svmsmoset function	Still runs	svmtrain and statset functions	svmsmoset is not recommended. Use svmtrain and statset instead.

Version 4.0 (R2011b) Bioinformatics Toolbox Software

Version 3.6 (R2010b) Bioinformatics Toolbox Software

© 1984-2012- The MathWorks, Inc.    -   Site Help   -   Patents   -   Trademarks   -   Privacy Policy   -   Preventing Piracy   -   RSS