Possibly use an mRNA detection method by looking at what is present in Plasma/serum. One method includes using RNAovation technologies from NUgen and deep sequencing and blast the reads then use a program like MEGAN to help identify and plot the reads detected on an alignment tree. 

http://ab.inf.uni-tuebingen.de/software/megan5/

I agree with Konstantiinidis: direct method is blood culure method. In addition to this and before the bacterial ID, you can do a Gram stain directly from the bottle after the growth signal,  In this way you can inform early the clinicians if the sepsi is from Gram+, Gram- or  fungi.

hi...we can detect bacteria by the measurement  of  genomic DNA - its a measure of a single molecules of DNA in arrays.

i feels, this method is a highly sensitive and even the sample volume also  very less just like 25 to 50microliter .its also the biotinylated detection and using the paramagnetic beads and are detected in a single molecule of array.

and also agree the above mentioned methods ,but i feels that those methods are  not a highly  sensitve and and the count for grading also ..just like approximation with a reciprocal calculation based on the dilution used.

and also the conventional fluoresence microscopy is very fast and highly sensitive.just 5 min.enough.its done by a laser scanning method.in this method individual bacteria can be stained and detected by laser..

the NuGen RNAovation based technologies are pretty sensitive but its still "PCR" based and therefore comes with some bias, not as much as PCR but still somewhat bias. Is anyone using 3rd gen sequencing (single molecule sequencing)? I'm interested in what 3rd gen could offer in terms of microbiome and  virome analyses. 

Direct detection of microbes in clinical samples is always a question for clinical microbiologists. From one site we must be aware about Koch’s postulates and confirm the relationship of a microbe with symptoms: the microorganism must be isolated from a diseased organism and grown in pure culture.

From the other site, if we need to diagnose a patient, we can apply algorithm which can connect “symptoms” with a “cause”. For such reason every method wich provides sensitive and specific discrimination criteria is useful and can be recommended:

1. immunomethods for detection of specific antigens (very quick and not expensive, if you used latex tests you would not even need special equipment)

2. PCR (and other molecular) completely different approach, you need special and sophisticated equipment, more time, more people but the sensitivity is much more robust.

3. Biochemistry – for detection of specific metabolites in cultures, commonly used

4. MALDI-TOF mass spectrometry, a revolutionary tool in microbiology. Using this method you can rapidly identify many bacteria strains, including organisms that are difficult to culture or new emerging strains. They also can be used as epidemiological tools to follow disease transmission

https://www.researchgate.net/publication/49660293_MALDI-TOF-mass_spectrometry_applications_in_clinical_microbiology

I hope that it will help!

Article MALDI-TOF-mass spectrometry applications in clinical microbiology

Completely in agreement with Ewe!

I AGREE WITH EWE.  Blood culture is the basic test which involve aerobic and anaerobic bacteria.  But there are specific tests too which can run by depending on sensitive and specific criteria.

comparing stool sample with blood sample...........hummm........... if your aim is to detect bacteria in blood, standard guidelines are available. You can find it with any clinical microbiology laboratory. But tools will be different with stool. .....

Easiest method...if present is the simplest...Culture using  blood culture bottle. When the concentration of growth in the specialized media is high enough the media is plated and a gram stain is performed. Micro organisms in blood cultures are text book perfect either-gram negative or gram positive rods or cocci  in pairs, clusters or chains. Following incubation and isolation the organism ID and sensitivities are performed  on an analyzer such as a MicroScan.

For direct from blood the challenge is achieving sensitivity. If the required sensitivity is 1 colony forming unit per ml of blood then you will need some way to concentrate the bacteria from 10 ml of blood. 1 cfu may not be detectable by PCR due to stochastic effects, also extraction etc will have to be optimized to reduce bacterial nucleic acid loss.

Blood culture has a high failure rate mostly due to presence of antibiotics in the blood or the inability of the bacteria to be grown with the culture conditions. And if it is a polymicrobial infection you may preferentially select for one of the many organisms by culture. 

As previously remembered by Ewa Stepien, we need consider the Koch's postulates. So, finding bacterial products (nucleic acids or proteins) has different meaning compared to finding living bacteria.